big five personality traits research paper

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

• View all journals
• Explore content
• About the journal
• Publish with us
• Sign up for alerts
• Open access
• Published: 22 May 2020

Assessing the Big Five personality traits using real-life static facial images

• Alexander Kachur   ORCID: orcid.org/0000-0003-1165-2672 1 ,
• Evgeny Osin   ORCID: orcid.org/0000-0003-3330-5647 2 ,
• Denis Davydov   ORCID: orcid.org/0000-0003-3747-7403 3 ,
• Konstantin Shutilov 4 &
• Alexey Novokshonov 4  

Scientific Reports volume  10 , Article number:  8487 ( 2020 ) Cite this article

131k Accesses

50 Citations

350 Altmetric

Metrics details

• Computer science
• Human behaviour

There is ample evidence that morphological and social cues in a human face provide signals of human personality and behaviour. Previous studies have discovered associations between the features of artificial composite facial images and attributions of personality traits by human experts. We present new findings demonstrating the statistically significant prediction of a wider set of personality features (all the Big Five personality traits) for both men and women using real-life static facial images. Volunteer participants (N = 12,447) provided their face photographs (31,367 images) and completed a self-report measure of the Big Five traits. We trained a cascade of artificial neural networks (ANNs) on a large labelled dataset to predict self-reported Big Five scores. The highest correlations between observed and predicted personality scores were found for conscientiousness (0.360 for men and 0.335 for women) and the mean effect size was 0.243, exceeding the results obtained in prior studies using ‘selfies’. The findings strongly support the possibility of predicting multidimensional personality profiles from static facial images using ANNs trained on large labelled datasets. Future research could investigate the relative contribution of morphological features of the face and other characteristics of facial images to predicting personality.

Similar content being viewed by others

Facial recognition technology can expose political orientation from naturalistic facial images

Individual differences and the multidimensional nature of face perception

Professional actors demonstrate variability, not stereotypical expressions, when portraying emotional states in photographs

Introduction.

A growing number of studies have linked facial images to personality. It has been established that humans are able to perceive certain personality traits from each other’s faces with some degree of accuracy 1 , 2 , 3 , 4 . In addition to emotional expressions and other nonverbal behaviours conveying information about one’s psychological processes through the face, research has found that valid inferences about personality characteristics can even be made based on static images of the face with a neutral expression 5 , 6 , 7 . These findings suggest that people may use signals from each other’s faces to adjust the ways they communicate, depending on the emotional reactions and perceived personality of the interlocutor. Such signals must be fairly informative and sufficiently repetitive for recipients to take advantage of the information being conveyed 8 .

Studies focusing on the objective characteristics of human faces have found some associations between facial morphology and personality features. For instance, facial symmetry predicts extraversion 9 . Another widely studied indicator is the facial width to height ratio (fWHR), which has been linked to various traits, such as achievement striving 10 , deception 11 , dominance 12 , aggressiveness 13 , 14 , 15 , 16 , and risk-taking 17 . The fWHR can be detected with high reliability irrespective of facial hair. The accuracy of fWHR-based judgements suggests that the human perceptual system may have evolved to be sensitive to static facial features, such as the relative face width 18 .

There are several theoretical reasons to expect associations between facial images and personality. First, genetic background contributes to both face and personality. Genetic correlates of craniofacial characteristics have been discovered both in clinical contexts 19 , 20 and in non-clinical populations 21 . In addition to shaping the face, genes also play a role in the development of various personality traits, such as risky behaviour 22 , 23 , 24 , and the contribution of genes to some traits exceeds the contribution of environmental factors 25 . For the Big Five traits, heritability coefficients reflecting the proportion of variance that can be attributed to genetic factors typically lie in the 0.30–0.60 range 26 , 27 . From an evolutionary perspective, these associations can be expected to have emerged by means of sexual selection. Recent studies have argued that some static facial features, such as the supraorbital region, may have evolved as a means of social communication 28 and that facial attractiveness signalling valuable personality characteristics is associated with mating success 29 .

Second, there is some evidence showing that pre- and postnatal hormones affect both facial shape and personality. For instance, the face is a visible indicator of the levels of sex hormones, such as testosterone and oestrogen, which affect the formation of skull bones and the fWHR 30 , 31 , 32 . Given that prenatal and postnatal sex hormone levels do influence behaviour, facial features may correlate with hormonally driven personality characteristics, such as aggressiveness 33 , competitiveness, and dominance, at least for men 34 , 35 . Thus, in addition to genes, the associations of facial features with behavioural tendencies may also be explained by androgens and potentially other hormones affecting both face and behaviour.

Third, the perception of one’s facial features by oneself and by others influences one’s subsequent behaviour and personality 36 . Just as the perceived ‘cleverness’ of an individual may lead to higher educational attainment 37 , prejudice associated with the shape of one’s face may lead to the development of maladaptive personality characteristics (i.e., the ‘Quasimodo complex’ 38 ). The associations between appearance and personality over the lifespan have been explored in longitudinal observational studies, providing evidence of ‘self-fulfilling prophecy’-type and ‘self-defeating prophecy’-type effects 39 .

Fourth and finally, some personality traits are associated with habitual patterns of emotionally expressive behaviour. Habitual emotional expressions may shape the static features of the face, leading to the formation of wrinkles and/or the development of facial muscles.

Existing studies have revealed the links between objective facial picture cues and general personality traits based on the Five-Factor Model or the Big Five (BF) model of personality 40 . However, a quick glance at the sizes of the effects found in these studies (summarized in Table  1 ) reveals much controversy. The results appear to be inconsistent across studies and hardly replicable 41 . These inconsistencies may result from the use of small samples of stimulus faces, as well as from the vast differences in methodologies. Stronger effect sizes are typically found in studies using composite facial images derived from groups of individuals with high and low scores on each of the Big Five dimensions 6 , 7 , 8 . Naturally, the task of identifying traits using artificial images comprised of contrasting pairs with all other individual features eliminated or held constant appears to be relatively easy. This is in contrast to realistic situations, where faces of individuals reflect a full range of continuous personality characteristics embedded in a variety of individual facial features.

Studies relying on photographic images of individual faces, either artificially manipulated 2 , 42 or realistic, tend to yield more modest effects. It appears that studies using realistic photographs made in controlled conditions (neutral expression, looking straight at the camera, consistent posture, lighting, and distance to the camera, no glasses, no jewellery, no make-up, etc.) produce stronger effects than studies using ‘selfies’ 25 . Unfortunately, differences in the methodologies make it hard to hypothesize whether the diversity of these findings is explained by variance in image quality, image background, or the prediction models used.

Research into the links between facial picture cues and personality traits faces several challenges. First, the number of specific facial features is very large, and some of them are hard to quantify. Second, the effects of isolated facial features are generally weak and only become statistically noticeable in large samples. Third, the associations between objective facial features and personality traits might be interactive and nonlinear. Finally, studies using real-life photographs confront an additional challenge in that the very characteristics of the images (e.g., the angle of the head, facial expression, makeup, hairstyle, facial hair style, etc.) are based on the subjects’ choices, which are potentially influenced by personality; after all, one of the principal reasons why people make and share their photographs is to signal to others what kind of person they are. The task of isolating the contribution of each variable out of the multitude of these individual variables appears to be hardly feasible. Instead, recent studies in the field have tended to rely on a holistic approach, investigating the subjective perception of personality based on integral facial images.

The holistic approach aims to mimic the mechanisms of human perception of the face and the ways in which people make judgements about each other’s personality. This approach is supported by studies of human face perception, showing that faces are perceived and encoded in a holistic manner by the human brain 43 , 44 , 45 , 46 . Put differently, when people identify others, they consider individual facial features (such as a person’s eyes, nose, and mouth) in concert as a single entity rather than as independent pieces of information 47 , 48 , 49 , 50 . Similar to facial identification, personality judgements involve the extraction of invariant facial markers associated with relatively stable characteristics of an individual’s behaviour. Existing evidence suggests that various social judgements might be based on a common visual representational system involving the holistic processing of visual information 51 , 52 . Thus, even though the associations between isolated facial features and personality characteristics sought by ancient physiognomists have emerged to be weak, contradictory or even non-existent, the holistic approach to understanding the face-personality links appears to be more promising.

An additional challenge faced by studies seeking to reveal the face-personality links is constituted by the inconsistency of the evaluations of personality traits by human raters. As a result, a fairly large number of human raters is required to obtain reliable estimates of personality traits for each photograph. In contrast, recent attempts at using machine learning algorithms have suggested that artificial intelligence may outperform individual human raters. For instance, S. Hu and colleagues 40 used the composite partial least squares component approach to analyse dense 3D facial images obtained in controlled conditions and found significant associations with personality traits (stronger for men than for women).

A similar approach can be implemented using advanced machine learning algorithms, such as artificial neural networks (ANNs), which can extract and process significant features in a holistic manner. The recent applications of ANNs to the analysis of human faces, body postures, and behaviours with the purpose of inferring apparent personality traits 53 , 54 indicate that this approach leads to a higher accuracy of prediction compared to individual human raters. The main difficulty of the ANN approach is the need for large labelled training datasets that are difficult to obtain in laboratory settings. However, ANNs do not require high-quality photographs taken in controlled conditions and can potentially be trained using real-life photographs provided that the dataset is large enough. The interpretation of findings in such studies needs to acknowledge that a real-life photograph, especially one chosen by a study participant, can be viewed as a holistic behavioural act, which may potentially contain other cues to the subjects’ personality in addition to static facial features (e.g., lighting, hairstyle, head angle, picture quality, etc.).

The purpose of the current study was to investigate the associations of facial picture cues with self-reported Big Five personality traits by training a cascade of ANNs to predict personality traits from static facial images. The general hypothesis is that a real-life photograph contains cues about personality that can be extracted using machine learning. Due to the vast diversity of findings concerning the prediction accuracy of different traits across previous studies, we did not set a priori hypotheses about differences in prediction accuracy across traits.

Prediction accuracy

We used data from the test dataset containing predicted scores for 3,137 images associated with 1,245 individuals. To determine whether the variance in the predicted scores was associated with differences across images or across individuals, we calculated the intraclass correlation coefficients (ICCs) presented in Table  2 . The between-individual proportion of variance in the predicted scores ranged from 79 to 88% for different traits, indicating a general consistency of predicted scores for different photographs of the same individual. We derived the individual scores used in all subsequent analyses as the simple averages of the predicted scores for all images provided by each participant.

The correlation coefficients between the self-report test scores and the scores predicted by the ANN ranged from 0.14 to 0.36. The associations were strongest for conscientiousness and weakest for openness. Extraversion and neuroticism were significantly better predicted for women than for men (based on the z test). We also compared the prediction accuracy within each gender using Steiger’s test for dependent sample correlation coefficients. For men, conscientiousness was predicted more accurately than the other four traits (the differences among the latter were not statistically significant). For women, conscientiousness was predicted more accurately, and openness was predicted less accurately compared to the three other traits.

The mean absolute error (MAE) of prediction ranged between 0.89 and 1.04 standard deviations. We did not find any associations between the number of photographs and prediction error.

Trait intercorrelations

The structure of the correlations between the scales was generally similar for the observed test scores and the predicted values, but some coefficients differed significantly (based on the z test) (see Table  3 ). Most notably, predicted openness was more strongly associated with conscientiousness (negatively) and extraversion (positively), whereas its association with agreeableness was negative rather than positive. The associations of predicted agreeableness with conscientiousness and neuroticism were stronger than those between the respective observed scores. In women, predicted neuroticism demonstrated a stronger inverse association with conscientiousness and a stronger positive association with openness. In men, predicted neuroticism was less strongly associated with extraversion than its observed counterpart.

To illustrate the findings, we created composite images using Abrosoft FantaMorph 5 by averaging the uploaded images across contrast groups of 100 individuals with the highest and the lowest test scores on each trait. The resulting morphed images in which individual features are eliminated are presented in Fig.  1 .

Composite facial images morphed across contrast groups of 100 individuals for each Big Five trait.

This study presents new evidence confirming that human personality is related to individual facial appearance. We expected that machine learning (in our case, artificial neural networks) could reveal multidimensional personality profiles based on static morphological facial features. We circumvented the reliability limitations of human raters by developing a neural network and training it on a large dataset labelled with self-reported Big Five traits.

We expected that personality traits would be reflected in the whole facial image rather than in its isolated features. Based on this expectation, we developed a novel two-tier machine learning algorithm to encode the invariant facial features as a vector in a 128-dimensional space that was used to predict the BF traits by means of a multilayer perceptron. Although studies using real-life photographs do not require strict experimental conditions, we had to undertake a series of additional organizational and technological steps to ensure consistent facial image characteristics and quality.

Our results demonstrate that real-life photographs taken in uncontrolled conditions can be used to predict personality traits using complex computer vision algorithms. This finding is in contrast to previous studies that mostly relied on high-quality facial images taken in controlled settings. The accuracy of prediction that we obtained exceeds that in the findings of prior studies that used realistic individual photographs taken in uncontrolled conditions (e.g., selfies 55 ). The advantage of our methodology is that it is relatively simple (e.g., it does not rely on 3D scanners or 3D facial landmark maps) and can be easily implemented using a desktop computer with a stock graphics accelerator.

In the present study, conscientiousness emerged to be more easily recognizable than the other four traits, which is consistent with some of the existing findings 7 , 40 . The weaker effects for extraversion and neuroticism found in our sample may be because these traits are associated with positive and negative emotional experiences, whereas we only aimed to use images with neutral or close to neutral emotional expressions. Finally, this appears to be the first study to achieve a significant prediction of openness to experience. Predictions of personality based on female faces appeared to be more reliable than those for male faces in our sample, in contrast to some previous studies 40 .

The BF factors are known to be non-orthogonal, and we paid attention to their intercorrelations in our study 56 , 57 . Various models have attempted to explain the BF using higher-order dimensions, such as stability and plasticity 58 or a single general factor of personality (GFP) 59 . We discovered that the intercorrelations of predicted factors tend to be stronger than the intercorrelations of self-report questionnaire scales used to train the model. This finding suggests a potential biological basis of GFP. However, the stronger intercorrelations of the predicted scores can be explained by consistent differences in picture quality (just as the correlations between the self-report scales can be explained by social desirability effects and other varieties of response bias 60 ). Clearly, additional research is needed to understand the context of this finding.

We believe that the present study, which did not involve any subjective human raters, constitutes solid evidence that all the Big Five traits are associated with facial cues that can be extracted using machine learning algorithms. However, despite having taken reasonable organizational and technical steps to exclude the potential confounds and focus on static facial features, we are still unable to claim that morphological features of the face explain all the personality-related image variance captured by the ANNs. Rather, we propose to see facial photographs taken by subjects themselves as complex behavioural acts that can be evaluated holistically and that may contain various other subtle personality cues in addition to static facial features.

The correlations reported above with a mean r = 0.243 can be viewed as modest; indeed, facial image-based personality assessment can hardly replace traditional personality measures. However, this effect size indicates that an ANN can make a correct guess about the relative standing of two randomly chosen individuals on a personality dimension in 58% of cases (as opposed to the 50% expected by chance) 61 . The effect sizes we observed are comparable with the meta-analytic estimates of correlations between self-reported and observer ratings of personality traits: the associations range from 0.30 to 0.49 when one’s personality is rated by close relatives or colleagues, but only from −0.01 to 0.29 when rated by strangers 62 . Thus, an artificial neural network relying on static facial images outperforms an average human rater who meets the target in person without any prior acquaintance. Given that partner personality and match between two personalities predict friendship formation 63 , long-term relationship satisfaction 64 , and the outcomes of dyadic interaction in unstructured settings 65 , the aid of artificial intelligence in making partner choices could help individuals to achieve more satisfying interaction outcomes.

There are a vast number of potential applications to be explored. The recognition of personality from real-life photos can be applied in a wide range of scenarios, complementing the traditional approaches to personality assessment in settings where speed is more important than accuracy. Applications may include suggesting best-fitting products or services to customers, proposing to individuals a best match in dyadic interaction settings (such as business negotiations, online teaching, etc.) or personalizing the human-computer interaction. Given that the practical value of any selection method is proportional to the number of decisions made and the size and variability of the pool of potential choices 66 , we believe that the applied potential of this technology can be easily revealed at a large scale, given its speed and low cost. Because the reliability and validity of self-report personality measures is not perfect, prediction could be further improved by supplementing these measures with peer ratings and objective behavioural indicators of personality traits.

The fact that conscientiousness was predicted better than the other traits for both men and women emerges as an interesting finding. From an evolutionary perspective, one would expect the traits most relevant for cooperation (conscientiousness and agreeableness) and social interaction (certain facets of extraversion and neuroticism, such as sociability, dominance, or hostility) to be reflected more readily in the human face. The results are generally in line with this idea, but they need to be replicated and extended by incorporating trait facets in future studies to provide support for this hypothesis.

Finally, although we tried to control the potential sources of confounds and errors by instructing the participants and by screening the photographs (based on angles, facial expressions, makeup, etc.), the present study is not without limitations. First, the real-life photographs we used could still carry a variety of subtle cues, such as makeup, angle, light facial expressions, and information related to all the other choices people make when they take and share their own photographs. These additional cues could say something about their personality, and the effects of all these variables are inseparable from those of static facial features, making it hard to draw any fundamental conclusions from the findings. However, studies using real-life photographs may have higher ecological validity compared to laboratory studies; our results are more likely to generalize to real-life situations where users of various services are asked to share self-pictures of their choice.

Another limitation pertains to a geographically bounded sample of individuals; our participants were mostly Caucasian and represented one cultural and age group (Russian-speaking adults). Future studies could replicate the effects using populations representing a more diverse variety of ethnic, cultural, and age groups. Studies relying on other sources of personality data (e.g., peer ratings or expert ratings), as well as wider sets of personality traits, could complement and extend the present findings.

Sample and procedure

The study was carried out in the Russian language. The participants were anonymous volunteers recruited through social network advertisements. They did not receive any financial remuneration but were provided with a free report on their Big Five personality traits. The data were collected online using a dedicated research website and a mobile application. The participants provided their informed consent, completed the questionnaires, reported their age and gender and were asked to upload their photographs. They were instructed to take or upload several photographs of their face looking directly at the camera with enough lighting, a neutral facial expression and no other people in the picture and without makeup.

Our goal was to obtain an out-of-sample validation dataset of 616 respondents of each gender to achieve 80% power for a minimum effect we considered to be of practical significance ( r  = 0.10 at p < 0.05), requiring a total of 6,160 participants of each gender in the combined dataset comprising the training and validation datasets. However, we aimed to gather more data because we expected that some online respondents might provide low-quality or non-genuine photographs and/or invalid questionnaire responses.

The initial sample included 25,202 participants who completed the questionnaire and uploaded a total of 77,346 photographs. The final combined dataset comprised 12,447 valid questionnaires and 31,367 associated photographs after the data screening procedures (below). The participants ranged in age from 18 to 60 (59.4% women, M = 27.61, SD = 12.73, and 40.6% men, M = 32.60, SD = 11.85). The dataset was split randomly into a training dataset (90%) and a test dataset (10%) used to validate the prediction model. The validation dataset included the responses of 505 men who provided 1224 facial images and 740 women who provided 1913 images. Due to the sexually dimorphic nature of facial features and certain personality traits (particularly extraversion 1 , 67 , 68 ), all the predictive models were trained and validated separately for male and female faces.

Ethical approval

The research was carried out in accordance with the Declaration of Helsinki. The study protocol was approved by the Research Ethics Committee of the Open University for the Humanities and Economics. We obtained the participants’ informed consent to use their data and photographs for research purposes and to publish generalized findings. The morphed group average images presented in the paper do not allow the identification of individuals. No information or images that could lead to the identification of study participants have been published.

Data screening

We excluded incomplete questionnaires (N = 3,035) and used indices of response consistency to screen out random responders 69 . To detect systematic careless responses, we used the modal response category count, maximum longstring (maximum number of identical responses given in sequence by participant), and inter-item standard deviation for each questionnaire. At this stage, we screened out the answers of individuals with zero standard deviations (N = 329) and a maximum longstring above 10 (N = 1,416). To detect random responses, we calculated the following person-fit indices: the person-total response profile correlation, the consistency of response profiles for the first and the second half of the questionnaire, the consistency of response profiles obtained based on equivalent groups of items, the number of polytomous Guttman errors, and the intraclass correlation of item responses within facets.

Next, we conducted a simulation by generating random sets of integers in the 1–5 range based on a normal distribution (µ = 3, σ = 1) and on the uniform distribution and calculating the same person-fit indices. For each distribution, we generated a training dataset and a test dataset, each comprised of 1,000 simulated responses and 1,000 real responses drawn randomly from the sample. Next, we ran a logistic regression model using simulated vs real responses as the outcome variable and chose an optimal cutoff point to minimize the misclassification error (using the R package optcutoff). The sensitivity value was 0.991 for the uniform distribution and 0.960 for the normal distribution, and the specificity values were 0.923 and 0.980, respectively. Finally, we applied the trained model to the full dataset and identified observations predicted as likely to be simulated based on either distribution (N = 1,618). The remaining sample of responses (N = 18,804) was used in the subsequent analyses.

Big Five measure

We used a modified Russian version of the 5PFQ questionnaire 70 , which is a 75-item measure of the Big Five model, with 15 items per trait grouped into five three-item facets. To confirm the structural validity of the questionnaire, we tested an exploratory structural equation (ESEM) model with target rotation in Mplus 8.2. The items were treated as ordered categorical variables using the WLSMV estimator, and facet variance was modelled by introducing correlated uniqueness values for the items comprising each facet.

The theoretical model showed a good fit to the data (χ 2  = 147854.68, df = 2335, p < 0.001; CFI = 0.931; RMSEA = 0.040 [90% CI: 0.040, 0.041]; SRMR = 0.024). All the items showed statistically significant loadings on their theoretically expected scales (λ ranged from 0.14 to 0.87, M = 0.51, SD = 0.17), and the absolute cross-loadings were reasonably low (M = 0.11, SD = 0.11). The distributions of the resulting scales were approximately normal (with skewness and kurtosis values within the [−1; 1] range). To assess the reliability of the scales, we calculated two internal consistency indices, namely, robust omega (using the R package coefficientalpha) and algebraic greatest lower bound (GLB) reliability (using the R package psych) 71 (see Table  4 ).

Image screening and pre-processing

The images (photographs and video frames) were subjected to a three-step screening procedure aimed at removing fake and low-quality images. First, images with no human faces or with more than one human face were detected by our computer vision (CV) algorithms and automatically removed. Second, celebrity images were identified and removed by means of a dedicated neural network trained on a celebrity photo dataset (CelebFaces Attributes Dataset (CelebA), N > 200,000) 72 that was additionally enriched with pictures of Russian celebrities. The model showed a 98.4% detection accuracy. Third, we performed a manual moderation of the remaining images to remove images with partially covered faces, those that were evidently photoshopped or any other fake images not detected by CV.

The images retained for subsequent processing were converted to single-channel 8-bit greyscale format using the OpenCV framework (opencv.org). Head position (pitch, yaw, roll) was measured using our own dedicated neural network (multilayer perceptron) trained on a sample of 8 000 images labelled by our team. The mean absolute error achieved on the test sample of 800 images was 2.78° for roll, 1.67° for pitch, and 2.34° for yaw. We used the head position data to retain the images with yaw and roll within the −30° to 30° range and pitch within the −15° to 15° range.

Next, we assessed emotional neutrality using the Microsoft Cognitive Services API on the Azure platform (score range: 0 to 1) and used 0.50 as a threshold criterion to remove emotionally expressive images. Finally, we applied the face and eye detection, alignment, resize, and crop functions available within the Dlib (dlib.net) open-source toolkit to arrive at a set of standardized 224 × 224 pixel images with eye pupils aligned to a standard position with an accuracy of 1 px. Images with low resolution that contained less than 60 pixels between the eyes, were excluded in the process.

The final photoset comprised 41,835 images. After the screened questionnaire responses and images were joined, we obtained a set of 12,447 valid Big Five questionnaires associated with 31,367 validated images (an average of 2.59 images per person for women and 2.42 for men).

Neural network architecture

First, we developed a computer vision neural network (NNCV) aiming to determine the invariant features of static facial images that distinguish one face from another but remain constant across different images of the same person. We aimed to choose a neural network architecture with a good feature space and resource-efficient learning, considering the limited hardware available to our research team. We chose a residual network architecture based on ResNet 73 (see Fig.  2 ).

Layer architecture of the computer vision neural network (NNCV) and the personality diagnostics neural network (NNPD).

This type of neural network was originally developed for image classification. We dropped the final layer from the original architecture and obtained a NNCV that takes a static monochrome image (224 × 224 pixels in size) and generates a vector of 128 32-bit dimensions describing unique facial features in the source image. As a measure of success, we calculated the Euclidean distance between the vectors generated from different images.

Using Internet search engines, we collected a training dataset of approximately 2 million openly available unlabelled real-life photos taken in uncontrolled conditions stratified by race, age and gender (using search engine queries such as ‘face photo’, ‘face pictures’, etc.). The training was conducted on a server equipped with four NVidia Titan accelerators. The trained neural network was validated on a dataset of 40,000 images belonging to 800 people, which was an out-of-sample part of the original dataset. The Euclidean distance threshold for the vectors belonging to the same person was 0.40 after the training was complete.

Finally, we trained a personality diagnostics neural network (NNPD), which was implemented as a multilayer perceptron (see Fig.  2 ). For that purpose, we used a training dataset (90% of the final sample) containing the questionnaire scores of 11,202 respondents and a total of 28,230 associated photographs. The NNPD takes the vector of the invariants obtained from NNCV as an input and predicts the Big Five personality traits as the output. The network was trained using the same hardware, and the training process took 9 days. The whole process was performed for male and female faces separately.

Data availability

The set of photographs is not made available because we did not solicit the consent of the study participants to publish the individual photographs. The test dataset with the observed and predicted Big Five scores is available from the openICPSR repository: https://doi.org/10.3886/E109082V1 .

Kramer, R. S. S., King, J. E. & Ward, R. Identifying personality from the static, nonexpressive face in humans and chimpanzees: Evidence of a shared system for signaling personality. Evol. Hum. Behav . https://doi.org/10.1016/j.evolhumbehav.2010.10.005 (2011).

Walker, M. & Vetter, T. Changing the personality of a face: Perceived big two and big five personality factors modeled in real photographs. J. Pers. Soc. Psychol. 110 , 609–624 (2016).

Article   Google Scholar  

Naumann, L. P., Vazire, S., Rentfrow, P. J. & Gosling, S. D. Personality Judgments Based on Physical Appearance. Personal. Soc. Psychol. Bull. 35 , 1661–1671 (2009).

Borkenau, P., Brecke, S., Möttig, C. & Paelecke, M. Extraversion is accurately perceived after a 50-ms exposure to a face. J. Res. Pers. 43 , 703–706 (2009).

Shevlin, M., Walker, S., Davies, M. N. O., Banyard, P. & Lewis, C. A. Can you judge a book by its cover? Evidence of self-stranger agreement on personality at zero acquaintance. Pers. Individ. Dif . https://doi.org/10.1016/S0191-8869(02)00356-2 (2003).

Penton-Voak, I. S., Pound, N., Little, A. C. & Perrett, D. I. Personality Judgments from Natural and Composite Facial Images: More Evidence For A “Kernel Of Truth” In Social Perception. Soc. Cogn. 24 , 607–640 (2006).

Little, A. C. & Perrett, D. I. Using composite images to assess accuracy in personality attribution to faces. Br. J. Psychol. 98 , 111–126 (2007).

Kramer, R. S. S. & Ward, R. Internal Facial Features are Signals of Personality and Health. Q. J. Exp. Psychol. 63 , 2273–2287 (2010).

Pound, N., Penton-Voak, I. S. & Brown, W. M. Facial symmetry is positively associated with self-reported extraversion. Pers. Individ. Dif. 43 , 1572–1582 (2007).

Lewis, G. J., Lefevre, C. E. & Bates, T. Facial width-to-height ratio predicts achievement drive in US presidents. Pers. Individ. Dif. 52 , 855–857 (2012).

Haselhuhn, M. P. & Wong, E. M. Bad to the bone: facial structure predicts unethical behaviour. Proc. R. Soc. B Biol. Sci. 279 , 571 LP–576 (2012).

Valentine, K. A., Li, N. P., Penke, L. & Perrett, D. I. Judging a Man by the Width of His Face: The Role of Facial Ratios and Dominance in Mate Choice at Speed-Dating Events. Psychol. Sci . 25 , (2014).

Carre, J. M. & McCormick, C. M. In your face: facial metrics predict aggressive behaviour in the laboratory and in varsity and professional hockey players. Proc. R. Soc. B Biol. Sci. 275 , 2651–2656 (2008).

Carré, J. M., McCormick, C. M. & Mondloch, C. J. Facial structure is a reliable cue of aggressive behavior: Research report. Psychol. Sci . https://doi.org/10.1111/j.1467-9280.2009.02423.x (2009).

Haselhuhn, M. P., Ormiston, M. E. & Wong, E. M. Men’s Facial Width-to-Height Ratio Predicts Aggression: A Meta-Analysis. PLoS One 10 , e0122637 (2015).

Lefevre, C. E., Etchells, P. J., Howell, E. C., Clark, A. P. & Penton-Voak, I. S. Facial width-to-height ratio predicts self-reported dominance and aggression in males and females, but a measure of masculinity does not. Biol. Lett . 10 , (2014).

Welker, K. M., Goetz, S. M. M. & Carré, J. M. Perceived and experimentally manipulated status moderates the relationship between facial structure and risk-taking. Evol. Hum. Behav . https://doi.org/10.1016/j.evolhumbehav.2015.03.006 (2015).

Geniole, S. N. & McCormick, C. M. Facing our ancestors: judgements of aggression are consistent and related to the facial width-to-height ratio in men irrespective of beards. Evol. Hum. Behav. 36 , 279–285 (2015).

Valentine, M. et al . Computer-Aided Recognition of Facial Attributes for Fetal Alcohol Spectrum Disorders. Pediatrics 140 , (2017).

Ferry, Q. et al . Diagnostically relevant facial gestalt information from ordinary photos. Elife 1–22 https://doi.org/10.7554/eLife.02020.001 (2014).

Claes, P. et al . Modeling 3D Facial Shape from DNA. PLoS Genet. 10 , e1004224 (2014).

Carpenter, J. P., Garcia, J. R. & Lum, J. K. Dopamine receptor genes predict risk preferences, time preferences, and related economic choices. J. Risk Uncertain. 42 , 233–261 (2011).

Dreber, A. et al . The 7R polymorphism in the dopamine receptor D4 gene (<em>DRD4</em>) is associated with financial risk taking in men. Evol. Hum. Behav. 30 , 85–92 (2009).

Bouchard, T. J. et al . Sources of human psychological differences: the Minnesota Study of Twins Reared Apart. Science (80-.). 250 , 223 LP–228 (1990).

Article   ADS   Google Scholar  

Livesley, W. J., Jang, K. L. & Vernon, P. A. Phenotypic and genetic structure of traits delineating personality disorder. Arch. Gen. Psychiatry https://doi.org/10.1001/archpsyc.55.10.941 (1998).

Bouchard, T. J. & Loehlin, J. C. Genes, evolution, and personality. Behavior Genetics https://doi.org/10.1023/A:1012294324713 (2001).

Vukasović, T. & Bratko, D. Heritability of personality: A meta-analysis of behavior genetic studies. Psychol. Bull. 141 , 769–785 (2015).

Godinho, R. M., Spikins, P. & O’Higgins, P. Supraorbital morphology and social dynamics in human evolution. Nat. Ecol. Evol . https://doi.org/10.1038/s41559-018-0528-0 (2018).

Rhodes, G., Simmons, L. W. & Peters, M. Attractiveness and sexual behavior: Does attractiveness enhance mating success? Evol. Hum. Behav . https://doi.org/10.1016/j.evolhumbehav.2004.08.014 (2005).

Lefevre, C. E., Lewis, G. J., Perrett, D. I. & Penke, L. Telling facial metrics: Facial width is associated with testosterone levels in men. Evol. Hum. Behav. 34 , 273–279 (2013).

Whitehouse, A. J. O. et al . Prenatal testosterone exposure is related to sexually dimorphic facial morphology in adulthood. Proceedings. Biol. Sci. 282 , 20151351 (2015).

Penton-Voak, I. S. & Chen, J. Y. High salivary testosterone is linked to masculine male facial appearance in humans. Evol. Hum. Behav . https://doi.org/10.1016/j.evolhumbehav.2004.04.003 (2004).

Carré, J. M. & Archer, J. Testosterone and human behavior: the role of individual and contextual variables. Curr. Opin. Psychol. 19 , 149–153 (2018).

Swaddle, J. P. & Reierson, G. W. Testosterone increases perceived dominance but not attractiveness in human males. Proc. R. Soc. B Biol. Sci . https://doi.org/10.1098/rspb.2002.2165 (2002).

Eisenegger, C., Kumsta, R., Naef, M., Gromoll, J. & Heinrichs, M. Testosterone and androgen receptor gene polymorphism are associated with confidence and competitiveness in men. Horm. Behav. 92 , 93–102 (2017).

Article   CAS   Google Scholar  

Kaplan, H. B. Social Psychology of Self-Referent Behavior . https://doi.org/10.1007/978-1-4899-2233-5 . (Springer US, 1986).

Rosenthal, R. & Jacobson, L. Pygmalion in the classroom. Urban Rev . https://doi.org/10.1007/BF02322211 (1968).

Masters, F. W. & Greaves, D. C. The Quasimodo complex. Br. J. Plast. Surg . 204–210 (1967).

Zebrowitz, L. A., Collins, M. A. & Dutta, R. The Relationship between Appearance and Personality Across the Life Span. Personal. Soc. Psychol. Bull. 24 , 736–749 (1998).

Hu, S. et al . Signatures of personality on dense 3D facial images. Sci. Rep. 7 , 73 (2017).

Kosinski, M. Facial Width-to-Height Ratio Does Not Predict Self-Reported Behavioral Tendencies. Psychol. Sci. 28 , 1675–1682 (2017).

Walker, M., Schönborn, S., Greifeneder, R. & Vetter, T. The basel face database: A validated set of photographs reflecting systematic differences in big two and big five personality dimensions. PLoS One 13 , (2018).

Goffaux, V. & Rossion, B. Faces are ‘spatial’ - Holistic face perception is supported by low spatial frequencies. J. Exp. Psychol. Hum. Percept. Perform . https://doi.org/10.1037/0096-1523.32.4.1023 (2006).

Schiltz, C. & Rossion, B. Faces are represented holistically in the human occipito-temporal cortex. Neuroimage https://doi.org/10.1016/j.neuroimage.2006.05.037 (2006).

Van Belle, G., De Graef, P., Verfaillie, K., Busigny, T. & Rossion, B. Whole not hole: Expert face recognition requires holistic perception. Neuropsychologia https://doi.org/10.1016/j.neuropsychologia.2010.04.034 (2010).

Quadflieg, S., Todorov, A., Laguesse, R. & Rossion, B. Normal face-based judgements of social characteristics despite severely impaired holistic face processing. Vis. cogn. 20 , 865–882 (2012).

McKone, E. Isolating the Special Component of Face Recognition: Peripheral Identification and a Mooney Face. J. Exp. Psychol. Learn. Mem. Cogn . https://doi.org/10.1037/0278-7393.30.1.181 (2004).

Sergent, J. An investigation into component and configural processes underlying face perception. Br. J. Psychol . https://doi.org/10.1111/j.2044-8295.1984.tb01895.x (1984).

Tanaka, J. W. & Farah, M. J. Parts and Wholes in Face Recognition. Q. J. Exp. Psychol. Sect. A https://doi.org/10.1080/14640749308401045 (1993).

Young, A. W., Hellawell, D. & Hay, D. C. Configurational information in face perception. Perception https://doi.org/10.1068/p160747n (2013).

Calder, A. J. & Young, A. W. Understanding the recognition of facial identity and facial expression. Nature Reviews Neuroscience https://doi.org/10.1038/nrn1724 (2005).

Todorov, A., Loehr, V. & Oosterhof, N. N. The obligatory nature of holistic processing of faces in social judgments. Perception https://doi.org/10.1068/p6501 (2010).

Junior, J. C. S. J. et al . First Impressions: A Survey on Computer Vision-Based Apparent Personality Trait Analysis. (2018).

Wang, Y. & Kosinski, M. Deep neural networks are more accurate than humans at detecting sexual orientation from facial images. J. Pers. Soc. Psychol. 114 , 246–257 (2018).

Qiu, L., Lu, J., Yang, S., Qu, W. & Zhu, T. What does your selfie say about you? Comput. Human Behav. 52 , 443–449 (2015).

Digman, J. M. Higher order factors of the Big Five. J.Pers.Soc.Psychol . https://doi.org/10.1037/0022-3514.73.6.1246 (1997).

Musek, J. A general factor of personality: Evidence for the Big One in the five-factor model. J. Res. Pers . https://doi.org/10.1016/j.jrp.2007.02.003 (2007).

DeYoung, C. G. Higher-order factors of the Big Five in a multi-informant sample. J. Pers. Soc. Psychol. 91 , 1138–1151 (2006).

Rushton, J. P. & Irwing, P. A General Factor of Personality (GFP) from two meta-analyses of the Big Five: Digman (1997) and Mount, Barrick, Scullen, and Rounds (2005). Pers. Individ. Dif. 45 , 679–683 (2008).

Wood, D., Gardner, M. H. & Harms, P. D. How functionalist and process approaches to behavior can explain trait covariation. Psychol. Rev. 122 , 84–111 (2015).

Dunlap, W. P. Generalizing the Common Language Effect Size indicator to bivariate normal correlations. Psych. Bull. 116 , 509–511 (1994).

Connolly, J. J., Kavanagh, E. J. & Viswesvaran, C. The convergent validity between self and observer ratings of personality: A meta-analytic review. Int. J. of Selection and Assessment. 15 , 110–117 (2007).

Harris, K. & Vazire, S. On friendship development and the Big Five personality traits. Soc. and Pers. Psychol. Compass. 10 , 647–667 (2016).

Weidmann, R., Schönbrodt, F. D., Ledermann, T. & Grob, A. Concurrent and longitudinal dyadic polynomial regression analyses of Big Five traits and relationship satisfaction: Does similarity matter? J. Res. in Personality. 70 , 6–15 (2017).

Cuperman, R. & Ickes, W. Big Five predictors of behavior and perceptions in initial dyadic interactions: Personality similarity helps extraverts and introverts, but hurts “disagreeables”. J. of Pers. and Soc. Psychol. 97 , 667–684 (2009).

Schmidt, F. L. & Hunter, J. E. The validity and utility of selection methods in personnel psychology: Practical and theoretical implications of 85 years of research findings. Psychol. Bull. 124 , 262–274 (1998).

Brown, M. & Sacco, D. F. Unrestricted sociosexuality predicts preferences for extraverted male faces. Pers. Individ. Dif. 108 , 123–127 (2017).

Lukaszewski, A. W. & Roney, J. R. The origins of extraversion: joint effects of facultative calibration and genetic polymorphism. Pers. Soc. Psychol. Bull. 37 , 409–21 (2011).

Curran, P. G. Methods for the detection of carelessly invalid responses in survey data. J. Exp. Soc. Psychol. 66 , 4–19 (2016).

Khromov, A. B. The five-factor questionnaire of personality [Pjatifaktornyj oprosnik lichnosti]. In Rus. (Kurgan State University, 2000).

Trizano-Hermosilla, I. & Alvarado, J. M. Best alternatives to Cronbach’s alpha reliability in realistic conditions: Congeneric and asymmetrical measurements. Front. Psychol . https://doi.org/10.3389/fpsyg.2016.00769 (2016).

Liu, Z., Luo, P., Wang, X. & Tang, X. Deep Learning Face Attributes in the Wild. in 2015 IEEE International Conference on Computer Vision (ICCV) 3730–3738 https://doi.org/10.1109/ICCV.2015.425 (IEEE, 2015).

He, K., Zhang, X., Ren, S. & Sun, J. Deep Residual Learning for Image Recognition. in 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) 770–778 https://doi.org/10.1109/CVPR.2016.90 (IEEE, 2016).

Download references

Acknowledgements

We appreciate the assistance of Oleg Poznyakov, who organized the data collection, and we are grateful to the anonymous peer reviewers for their detailed and insightful feedback.

Contributions

A.K., E.O., D.D. and A.N. designed the study. K.S. and A.K. designed the ML algorithms and trained the ANN. A.N. contributed to the data collection. A.K., K.S. and D.D. contributed to data pre-processing. E.O., D.D. and A.K. analysed the data, contributed to the main body of the manuscript, and revised the text. A.K. prepared Figs. 1 and 2. All the authors contributed to the final version of the manuscript.

Corresponding authors

Correspondence to Alexander Kachur or Evgeny Osin .

Ethics declarations

Competing interests.

A.K., K.S. and A.N. were employed by the company that provided the datasets for the research. E.O. and D.D. declare no competing interests.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary information., rights and permissions.

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/ .

Reprints and permissions

About this article

Cite this article.

Kachur, A., Osin, E., Davydov, D. et al. Assessing the Big Five personality traits using real-life static facial images. Sci Rep 10 , 8487 (2020). https://doi.org/10.1038/s41598-020-65358-6

Download citation

Received : 12 April 2019

Accepted : 28 April 2020

Published : 22 May 2020

DOI : https://doi.org/10.1038/s41598-020-65358-6

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

This article is cited by

Personality of organizational social media accounts and its relationship with characteristics of their photos: analyses of startups’ instagram photos.

• Yunhwan Kim

BMC Psychology (2024)

Seeing the darkness: identifying the Dark Triad from emotionally neutral faces

• Danielle Haroun
• Yaarit Amram
• Joseph Glicksohn

Current Psychology (2024)

A Richer Vocabulary of Chinese Personality Traits: Leveraging Word Embedding Technology for Mining Personality Descriptors

• Yigang Ding
• Feijun Zheng

Journal of Psycholinguistic Research (2024)

Unravelling the many facets of human cooperation in an experimental study

• Victoria V. Rostovtseva
• Mikael Puurtinen
• Franz J. Weissing

Scientific Reports (2023)

Using deep learning to predict ideology from facial photographs: expressions, beauty, and extra-facial information

• Stig Hebbelstrup Rye Rasmussen
• Steven G. Ludeke
• Robert Klemmensen

By submitting a comment you agree to abide by our Terms and Community Guidelines . If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.

Quick links

• Explore articles by subject
• Guide to authors
• Editorial policies

Sign up for the Nature Briefing: AI and Robotics newsletter — what matters in AI and robotics research, free to your inbox weekly.

ORIGINAL RESEARCH article

The relationship between big five personality and social well-being of chinese residents: the mediating effect of social support.

• 1 School of Public Finance and Management, Yunnan University of Finance and Economics, Kunming, China
• 2 Tourism and Cultural Industry Research Institute, Yunnan University of Finance and Economics, Kunming, China
• 3 International College, National Institute of Development Administration, Bangkok, Thailand
• 4 National Centre for Borderlands Ethnic Studies in Southwest China at Yunnan University (NaCBES), Yunnan University, Kunming, China

Previous studies have noted that personality traits are important predictors of well-being, but how big five personality influences social well-being is still unknown. This study aims to examine the link between big five personality and five dimensions of social well-being in the Chinese cultural context and whether social support can play the mediating effect in the process. This study included 1,658 participants from different communities in China, and regression analyses were conducted. Results revealed that five personality traits were significantly related to overall social well-being; extraversion was significantly related to social integration; agreeableness was positively related to all five dimensions of social well-being; conscientiousness was positively related to social actualization, social coherence, and social contribution; neuroticism was negatively related to social integration, social acceptance, social actualization, and social coherence; openness was positively related to social integration, social acceptance, social coherence, and social contribution. Social support plays mediating roles in the relationships between extraversion/agreeableness/conscientiousness/neuroticism/openness and social well-being, respectively.

Introduction

Personality variables are strong predictors of well-being, a large body of research has explored the associations between big five personality and subjective well-being ( DeNeve and Cooper, 1998 ; Gutiérrez et al., 2005 ). Unfortunately, the psychological construct of well-being portrays adult well-being as a primarily private phenomenon largely neglecting individuals’ social lives ( Keyes, 2002 ; Hill et al., 2012 ). Individuals are embedded in social structures and communities; as such, it is necessary to evaluate one’s circumstance and functioning in a society; more attention needs to be devoted on the topic of social well-being ( Keyes, 1998 ). Previous studies focused on the social well-being from the perspective of interpersonal factors, such as sense of community ( Sohi et al., 2017 ), and civic engagement ( Albanesi et al., 2010 ). However, less work has examined social well-being from the level of the individual ( Keyes and Shapiro, 2004 ).

Although there are few studies focusing on the relationship between five personality traits and social well-being ( Hill et al., 2012 ; Joshanloo et al., 2012 ), their data come from United States or Iran; Chinese cultural background has been conducted to a lesser extent. Different countries have different cultural traditions. Personality is created through the process of enculturation ( Hofstede and McCrae, 2004 ). The interplay of personality and cultural factors was found to predict residents’ well-being significantly ( Diener and Diener, 1995 ). Confucius culture has embedded itself in the daily life of the Chinese, however, studies about the relationship between personality and social well-being under the context of Chinese culture are largely overlooked.

In addition, present studies ( Hill et al., 2012 ; Joshanloo et al., 2012 ) examine only the direct effect of personality on social well-being. The mechanism between big five personality and five dimensions of social well-being has been neglected. Additionally, social support can help individuals protect against the health consequences of life stress and increase their well-being ( Cobb, 1976 ; Siedlecki et al., 2014 ). Thus, following a social support perspective, the present study examined not only the relationship between five personality traits and domains of social well-being, but also whether social support can play a mediating effect in the relationship between big five personality and social well-being.

Literature Review and Hypothesis

Big five personality and social well-being.

The big five personality consists of five general traits: extraversion, neuroticism, openness, agreeableness, and conscientiousness ( John and Srivastava, 1999 ). Extraversion refers to the degree to which one is energetic, social, talkative, and gregarious. Agreeableness reflects the extent to which one is warm, caring, supportive, and cooperative and gets along well with others. Conscientiousness involves the extent to which one is well-organized, responsible, punctual, achievement-oriented, and dependable. Neuroticism means the degree to which one is worry, anxious, impulsive, and insecure. Openness reflects the degree to which one is imaginative, creative, curious, and broad-minded ( Barrick et al., 2001 ; Funder and Fast, 2010 ). Many scholars assessed personality under different culture context by a combined emic–etic approach ( John and Srivastava, 1999 ; Cheung et al., 2001 ). Even if there were researches that demonstrated several unique dimensions of personality under the Chinese culture ( Cheung et al., 2001 ; Cheung, 2004 ), the generalizability of the big five trait taxonomy in China is still confirmed ( Li and Chen, 2015 ; Minkov et al., 2019 ). Previous studies have consistently demonstrated that the big five are associated with subjective well-being ( DeNeve and Cooper, 1998 ; Gutiérrez et al., 2005 ), however, the findings are mixed under different cultural context. For instance, Ha and Kim (2013) found openness has a positive effect on subjective well-being in South Korea residents, whereas another study by Hayes and Joseph (2003) in England found that openness was not associated with each of the three measures of subjective well-being.

Culture variables can explain differences in mean levels of well-being ( Diener et al., 2003 ). With the uniqueness of Confucian cultural tradition and social setting, it is noteworthy to discuss the relationship between personality and well-being in Chinese cultural background, especially social-well-being.

Individuals are embedded in social structures. They need to face social challenges and evaluate their life quality and personal functioning by comparison to social criteria ( Keyes and Shapiro, 2004 ). However, the research about social well-being has been almost completely neglected in the hedonic and psychological well-being models ( Keyes, 2002 ; Joshanloo et al., 2012 ). Keyes (1998) proposed social well-being, which indicates to what degree individuals are functioning well in the social world they are embedded in. Social well-being can be described on multiple dimensions, including social integration, social contribution, social acceptance, social coherence, and social actualization. Social integration is the extent to which people feel commonality and connectedness to their neighborhood, community, and society. Social contribution refers to a value evaluation that one can provide to the society. Social acceptance entails a positive view of human nature and believes that people are kind. Social coherence refers to the perception of the quality and operation of the social world and reflects a belief that society is meaningful. Social actualization is the evolution of the potential and of society and includes a sense that social potentials can be realized through its institutions and citizens. In summary, social well-being emphasizes individuals’ perceptions of and attitudes toward the whole society. Prior studies have found the effect of sense of community ( Sohi et al., 2017 ), and social participation ( Albanesi et al., 2010 ) on social well-being, Also, some studies have shown the outcomes of social well-being, such as anxiety problems ( Keyes, 2005 ), general mental and physical health ( Zhang et al., 2011 ), and prosocial behaviors ( Keyes and Ryff, 1998 ). Personality traits and cultural factors are important predictors of well-being ( Diener et al., 2003 ). However, the only studies about personality and social well-being were conducted in Iran or United States. It is still not known whether the association would be similar in a different cultural context ( Hill et al., 2012 ; Joshanloo et al., 2012 ). For example, with the data from the MIDUS sample, Hill et al. (2012) found social well-being is positively related to extraversion, agreeableness, conscientiousness, emotional stability, and openness. In addition, previous studies did not test the correlation between five personality traits and five domains of social well-being entirely ( Joshanloo et al., 2012 ). Personality shapes many of the attitudes and behaviors that form Keyes’ different dimensions of social well-being. Thus, certain personalities would predict social well-being; for example, extraverted persons should be more socially integrated, whereas agreeable individuals should possess higher levels of social acceptance. Based on the above, we hypothesize the following:

Hypothesis 1 a : Extraversion is positively related to social well-being.

Hypothesis 1 b : Agreeableness is positively related to social well-being.

Hypothesis 1 c : Conscientiousness is positively related to social well-being.

Hypothesis 1 d : Neuroticism is negatively related to social well-being.

Hypothesis 1 e : Openness is positively related to social well-being.

The Mediating Effect of Social Support

Social support refers to individuals’ psychological or material resources from their own social networks that can assist them to cope with stressful challenges in daily lives ( Cohen, 2004 ). It comes from a variety of sources, such as friends, family, and significant others ( Taylor, 2011 ). Social support comprised both received and perceived social support ( Oh et al., 2014 ; Hartley and Coffee, 2019 ). However, many studies showed that perceived social support is more effective at predicting residents’ mental health than the received social support ( Cohen and Syme, 1985 ). Perceived social support indicates recipients’ perceptions concerning the general availability of support ( Sarason et al., 1990 ), which fosters a sense of social connectedness in a network and provides resources with which to overcome obstacles in their lives ( Lee et al., 2001 ; Chen, 2013 ). Social support theory emphasizes that social support is an important resource that can help individuals protect against life stress and increase their quality of lives ( Cobb, 1976 ; Cohen and Wills, 1985 ). Numerous studies have explored the associations between social support and well-being, including subjective well-being ( Brannan et al., 2013 ; Siedlecki et al., 2014 ) and psychological well-being ( Jasinskaja-Lahti et al., 2006 ; Wong et al., 2007 ). Although Inoue et al. (2015) found social support mediated the effect of team identification on community coherence, little research has addressed the effect of social support on social well-being. The benefits of social support come into play when individuals have to deal with social challenges and problems. Individuals with high level of social supports will better face social tasks ( Cox, 2000 ). Harmonious social relationships can help residents to satisfy their social needs, better understand, and be confident of the social world. Therefore, their social well-being will increase.

Personality traits are stable predictors of social support ( Swickert et al., 2010 ; Udayar et al., 2018 ; Barańczuk, 2019 ). Big five personality traits are found to be related to social support. Individuals with high levels of neuroticism report greater vulnerability to stress and negative affectivity, which could decrease the availability of social support ( Ayub, 2015 ). Individuals who score high on extraversion always seek social interactions and tend to be cheerful and friendly. The positive emotions could increase their social support ( Swickert et al., 2010 ). Individuals with high openness to experience are characterized by greater openness to emotions, appreciation of art and beauty, intellect, and liberalism. These characteristics would be significantly related to social support ( Barańczuk, 2019 ). Agreeableness characteristics, such as modesty, compliance, and trust, may facilitate individuals building a more extensive social support network ( Barańczuk, 2019 ). Conscientiousness are characterized by achievement-striving, self-discipline, orderliness, and dutifulness. These tendencies can help individuals better cope with life stress, so it is positively related to social support ( Ayub, 2015 ). Culture is an important moderator between big five personality traits and social support association, but it has been largely overlooked in previous studies ( Barańczuk, 2019 ). Therefore, studies about the relationship between five personality traits and social support under Chinese background are needed.

Previous studies discuss only the direct effect of personality on social well-being, but it remains unknown what mechanism(s) may explain this relation. Social support plays an important stress-buffering role when individuals are under high levels of life stress ( Cohen, 2004 ). Individuals with different levels of personality traits (extraversion, agreeableness, conscientiousness, neuroticism, openness) will form different types of social support network. Further, social support will help individuals cope with social challenges and increase their social well-being. Based on the above, we hypothesize the following:

Hypothesis 2 a : Social support mediates the relationship between extraversion and social well-being.

Hypothesis 2 b : Social support mediates the relationship between agreeableness and social well-being.

Hypothesis 2 c : Social support mediates the relationship between conscientiousness and social well-being.

Hypothesis 2 d : Social support mediates the relationship between neuroticism and social well-being.

Hypothesis 2 e : Social support mediates the relationship between openness and social well-being.

Materials and Methods

Participants and procedure.

Community residents from five different districts in Kunming, Yunnan Province, were selected as participants by stratified random sampling technique. Four hundred questionnaires were distributed to each district. Participants would complete the questionnaires in a face-to-face interaction with an enumerator who helped them to answer the questionnaire that was in paper format. When we administered the survey, we emphasized that the data were collected for research purposes. Participants were encouraged to answer all the questions honestly and were reminded that their responses would be anonymous. Upon completion of answering the questionnaire, participants received a small gift (e.g., tissue) as compensation for their participation. A total of 2,000 questionnaires were distributed, and 1,721 responded. After dropping incomplete and invalid data, 1,658 respondents remained. The final sample consisted of 932 females (56.2%) and 726 males (43.8%), aged 18–81 years (mean = 30.73 years, SD = 11.98 years).

Big Five Personality

The 44-item Big Five Inventory (BFI; John et al., 1991 ) was used to measure the five broad personality traits. All items were evaluated on a 5-point Likert scale, ranging from “strongly disagree” to “strongly agree.” Coefficient α reliabilities for the five trait scales in the present study were 0.707 for extraversion, 0.712 for agreeableness, 0.729 for conscientiousness, 0.706 for neuroticism, and 0.733 for openness. The Chinese version of BFI we used had been translated from English using common back-translation procedures ( Brislin, 1970 ; Li and Chen, 2015 ), and the validity had been conformed in previous studies ( Zhou, 2010 ; Li and Chen, 2015 ).

Social Support

Participants rated their social support from Chen and Yu (2019) using scales that ranged from 1 (strongly disagree) to 5 (strongly agree). The measure comprised three items, such as “It is easy for me to find someone to help when I meet with difficulties.” The entire survey demonstrated good reliability (α = 0.733).

Social Well-Being

Social well-being was measured through Keyes’s (1998) 15-item scale composed of five dimensions: social actualization, social integration, social acceptance, social contribution, and social coherence. Responses to this measure were assessed on a 5-point scale, from “strongly disagree” to “strongly agree.” An example of measure items was “I believe that people are kind.” The reliabilities of five dimensions were good (ranging from 0.702 to 0.725), and overall α reliability for the present sample was 0.791. Previous studies had confirmed the validity of social well-being measurement of Chinese version we used ( Miao and Wang, 2009 ; Chen and Yu, 2019 ; Chen et al., 2020 ).

The Common Method Bias Examination

As one of the main sources of measurement error, common method variance is a potential problem, which may be a threat to the validity of the conclusions. We tested for common method bias with a single-factor measurement model by combining all items into a single factor ( Podsakoff et al., 2003 ; Rhee et al., 2017 ). Results showed a poor model fit [Comparative Fit Index (CFI) = 0.763, Tucker-Lewis Index (TLI) = 0.695, Goodness-of-Fit Index (GFI) = 0.719, Root Mean square Residual (RMR) = 0.025, Root Mean Square Error of Approximation (RMSEA) = 0.109]. The above results suggested that there was no common method bias effect.

Descriptive Statistics and Correlations Between the Study Variables

There is no significant difference between the five different districts in Kunming. The correlation coefficients, means, and standard deviations are shown in Table 1 . All the big five personality traits were correlated significantly with social support and five domains of social well-being (expect agreeableness and social coherence). Extraversion, agreeableness, conscientiousness, and openness were correlated positively with domains of social well-being (expect agreeableness and social coherence) and social support, whereas neuroticism correlated negatively with domains of social well-being and social support.

Table 1. Correlations, means, and standard deviations of all study variables.

Regression Analyses

Statistical analyses were conducted with the Statistical Package for Social Sciences (SPSS, version 22.0). Based on preliminary analyses, multiple regression analyses were conducted to assess the relationship between the big five personality domains and dimensions of social well-being. Both gender and age were statistically controlled during the regression analysis, because there is evidence to show that social well-being likely increases with one’s age ( Chen and Li, 2014 ) and that men generally score higher on well-being than women do ( Miao and Wang, 2009 ). OLS regression was used to test the hypothesis. In each regression analysis, one social well-being dimension was entered as the dependent variable; gender, age, and all five personality domains were entered as potential predictors. Results of the regression analyses are presented in Table 2 . Five personality traits were significant predictors of overall social well-being. Extraversion (β = 0.052, p ≤ 0.05), agreeableness (β = 0.197, p ≤ 0.001), conscientiousness (β = 0.138, p ≤ 0.001), and openness (β = 0.156, p ≤ 0.001) are positively related to social well-being, whereas neuroticism (β = −0.171, p ≤ 0.001) is negatively related to social well-being. H1 a , H1 b , H1 c , H1 d , and H1 e are supported. Extraversion (β = 0.118, p ≤ 0.001), agreeableness (β = 0.162, p ≤ 0.001), neuroticism (β = −0.065, p ≤ 0.05), and openness (β = 0.086, p ≤ 0.001) were significant predictors of social integration. Agreeableness (β = 0.268, p ≤ 0.001), neuroticism (β = −0.102, p ≤ 0.001), and openness (β = 0.089, p ≤ 0.001) were significantly associated with social acceptance. Agreeableness (β = 0.168, p ≤ 0.001), conscientiousness (β = 0.111, p ≤ 0.001), and neuroticism (β = −0.110, p ≤ 0.001) predicted social actualization significantly. Agreeableness (β = −0.088, p ≤ 0.001), conscientiousness (β = 0.060, p ≤ 0.05), neuroticism (β = −0.241, p ≤ 0.001), and openness (β = 0.125, p ≤ 0.001) were found to be predicting social coherence. Agreeableness (β = 0.120, p ≤ 0.001), conscientiousness (β = 0.191, p ≤ 0.001), and openness (β = 0.164, p ≤ 0.001) were found to be predictors of social contribution.

Table 2. Results of regression analyses for five personality traits predicting dimensions of social well-being.

Mediation Analyses

Further, mediation analysis was performed to determine whether the effect of big five personality on social well-being was mediated by social support. Mediation analyses were conducted following the recommendations of Preacher and Hayes (2004) , using the PROCESS macro (version 3.0), developed by Hayes (2013) . The current study used 5,000 bootstrapped samples with a 95% confidence interval. The results of this analysis are shown in Table 3 . The results suggested five personality traits are related to social support significantly, and social support is positively related to social well-being. In addition, social support mediated the relationship between five personality traits and social well-being. H2 a , H2 b , H2 c , H2 d , and H2 e are supported.

Table 3. Summary of mediation analyses on five personality traits and social well-being (5,000 bootstraps).

Discussion and Conclusion

The results obtained from the survey of 1,658 Chinese residents demonstrated the effects of five personality traits on five dimensions of social well-being and the mediating role of social support in the associations between big five personality and social well-being.

Theoretical Contributions

Research on linkages between big five personality domains and five dimensions of social well-being conducted in China will likely contribute to the extant personality and well-being literature. First, this study provides empirical evidence about the relationship between big five personality and social well-being. The association between the big five personality and social well-being was evidenced in our study. However, our research also showed some inconsistencies with previous researches ( Joshanloo et al., 2012 ). From our results, extraversion was significantly related to social integration; agreeableness was positively related to all five dimensions of social well-being; conscientiousness was positively related to social actualization, social coherence, and social contribution; neuroticism was negatively related to social integration, social acceptance, social actualization, and social coherence; openness was positively related to social integration, social acceptance, social coherence, and social contribution. This inconsistency may be explained by the fact that the differences between Iran and China. For instance, Iran is a non-Arab Muslim country; the interactions in Iran are regulated partly by religious norms ( Joshanloo et al., 2012 ). In China, with the Reform and Opening, the way of thinking and behavior of Chinese are becoming more and more open and innovative ( Ma, 2013 ). The goal of community construction in China is to establish the autonomous system of community residents ( Fei, 2002 ). Community residents’ committee is an important organization of residents’ self-governing and self-service ( Sun, 2016 ). Thus, most community residents can participate in community management and satisfy their own service needs via residents’ committee, which will benefit residents’ life quality.

Second, the study highlights the effect of social support on social well-being. The existing literature has shown the relationship between social support and subjective well-being or psychological well-being ( Jasinskaja-Lahti et al., 2006 ; Brannan et al., 2013 ). Further, our study demonstrated social support is positively related to social well-being. Well-being is increasingly being associated with social and cultural relationships ( Helliwell and Putnam, 2004 ). Community in China is increasingly becoming a place for residents to integrate into urban society ( Chen et al., 2020 ). One of the most important responsibilities of the community is to achieve the society reconstruction ( Fei, 2002 ). Thus, during the development of community, the Chinese government was committed to improving the quality of community services, which may provide more opportunities for residents to get more social support. Individuals having high social support means they had selected and built large and effective social networks, which can help to overcome difficulties in lives. With the help from their social relations, they will give a high appraisal to their circumstances and functioning in society; their social well-being also increases.

Third, the mediating effects were found for social support for relation between extraversion/agreeableness/conscientiousness/neuroticism/openness and social well-being. This may contribute to the literature on the relationship between big five personality and social well-being ( Hill et al., 2012 ; Joshanloo et al., 2012 ). Previous studies neglected to examine the relationship and the mechanism between big five personality and social well-being from the perspective of the community. Community is an important place for residents’ daily activities. Individuals with different personality traits may build their social relations in different ways. Friends or family or neighbors around them may behave with different reactions. The different levels of social support will influence their evaluation of the social world, which may cause different levels of social well-being.

Practical Implications

Our study provides valuable insight into how individuals of different traits to improve their social well-being. Social support serves as a mediator in the relationship between big five personality and social well-being. The results also affirm the importance of social support that can enhance social well-being. When one’s psychological, social, and/or resource needs are met, one is likely to experience greater social support, which is important for their well-being. Therefore, it is possible for residents to promote social support. Individuals should spend more time participating in community public affairs or other social activities that could offer opportunities for them to establish meaningful relationship with neighbors or friends.

Limitations and Future Research

Despite these findings, our research is not without limitations. First, culture is an important factor that can influence both personality traits and well-being ( Diener et al., 2003 ; Hofstede and McCrae, 2004 ). Our study just discussed the mediating effect of social support between personality and social well-being. Future research should explore the effects of different cultural variables (such as power distance, collectivism/individualism etc.,). In addition, comparative studies among different countries or regions are needed. Second, the cross-sectional design means that no causal conclusions for the found relationship can be made. Consequently, future researches should adopt longitudinal or experimental design to ascertain the relationship. Third, social support has usually been classified into several specific forms, such as informational support, emotional support, perceived social support ( Taylor, 2011 ). In current study, we just regarded perceived social support as the mediating variable. So, future research should examine the effects of different forms of social support.

The research used a sample drawn from 1,658 Chinese residents to investigate the relationship between big five personality and social well-being and the mediating effect of social support in the relationship between big five personality and social well-being. Results of this study support previous studies that highlighted the relationship between big five personality and social support ( Swickert et al., 2010 ; Barańczuk, 2019 ). In addition, this study demonstrated the effects of five personality traits on five dimensions of social well-being. Lastly, the results demonstrated the mediating role of social support in the associations between extraversion/agreeableness/conscientiousness/neuroticism/open ness and social well-being, respectively.

Data Availability Statement

The raw data supporting the conclusion of this article will be made available by the authors, without undue reservation.

Ethics Statement

The studies involving human participants were reviewed and approved by Yunnan University of Finance and Economics. Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements.

Author Contributions

YY, YZ, and JL designed the research and wrote the manuscript. YY and YZ are co-first authors of the article. All authors planned and conducted the data collection. YZ, JZ, and DL analyzed the data and revised the manuscript. All authors listed have made direct and intellectual contribution to the article and approved the final version for publication.

This study was supported by the Chinese National Natural Science Fund (72064042), the Post-project of Chinese Ministry of Education (18JHQ080), the Philosophy and Social Science Research Project in Yunnan Province (QN202026), and the Science Research Fund of Yunnan Provincial Department of Education (2020J0384).

Conflict of Interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Albanesi, C., Cicognani, E., and Zani, B. (2010). Sense of community, civic engagement and social well-being in italian adolescents. J. Commun. Appl. Soc. Psychol. 17, 387–406. doi: 10.1002/casp.903

CrossRef Full Text | Google Scholar

Ayub, N. (2015). Predicting suicide ideation through intrapersonal and interpersonal factors: the interplay of Big-Five personality traits and social support. Pers. Ment. Health 9, 308–318. doi: 10.1002/pmh.1301

PubMed Abstract | CrossRef Full Text | Google Scholar

Barańczuk, U. (2019). The five factor model of personality and social support: a meta-analysis. J. Res. Pers. 81, 38–46. doi: 10.1016/j.jrp.2019.05.002

Barrick, M. R., Mount, M. K., and Judge, T. A. (2001). Personality and performance at the beginning of the new millennium: what do we know and where do we go next? Intern. J. Select. Assess. 9, 9–30. doi: 10.1111/1468-2389.00160

Brannan, D., Biswas-Diener, R., Mohr, C. D., Mortazavi, S., and Stein, N. (2013). Friends and family: a cross-cultural investigation of social support and subjective well-being among college students. J. Posit. Psychol. 8, 65–75. doi: 10.1080/17439760.2012.743573

Brislin, R. (1970). Back-translation for cross-cultural research. J. Cross Cult. Psychol. 1, 185–216. doi: 10.1177/135910457000100301

Chen, L. H. (2013). Gratitude and adolescent Athletes’ well-being: the multiple mediating roles of perceived social support from coaches and teammates. Soc. Indic. Res. 114, 273–285. doi: 10.1007/s11205-012-0145-2

Chen, Z., Liu, S., Yu, Y., Bwanali, T. R., and Douangdara, V. (2020). Community satisfaction, sense of community, and social well-being in China. Soc. Behav. Pers. Intern. J. 48:e8648.

Google Scholar

Chen, Z. X., and Li, Q. M. (2014). Verification of integrated model of well-being in different age and gender groups. Chin. J. Clin. Psychol. 5, 78–83.

Chen, Z. X., and Yu, Y. H. (2019). Urban social management, social support and social well-being: an empirical study of 2284 Chinese in 41 Cities, Northwest population. Ann. Rev. Psychol. 40, 44–56.

Cheung, F. M. (2004). Use of Western-and indigenously-developed personality tests in Asia. Appl. Psychol. Intern. Rev. 53, 173–191. doi: 10.1111/j.1464-0597.2004.00167.x

Cheung, F. M., Leung, K., Zhang, J. X., Sun, H. F., Gan, Y. Q., Song, W. Z., et al. (2001). Indigenous Chinese personality constructs: is the five-factor model complete? J. Cross Cult. Psychol. 32, 407–433. doi: 10.1177/0022022101032004003

Cobb, S. (1976). Social support as a moderator of life stress. Psychosom. Med. 38, 300–314. doi: 10.1097/00006842-197609000-00003

Cohen, S. (2004). Social relationships and health. Am. Psychol. 59, 676–684.

Cohen, S., and Syme, S. L. (1985). “Issues in the study and application of social support,” in Social Support and Health , eds S. Cohen and S. L. Syme (Orlando, FL: Academic Press), 3–22.

Cohen, S., and Wills, T. A. (1985). Stress, social support and the buffering hypothesis. Psychol. Bull. 98, 310–357. doi: 10.1037/0033-2909.98.2.310

Cox, E. (2000). Creating a more civil society: community level indicators of social capital. Just Policy J. Austr. Soc. Policy 19, 100–107.

DeNeve, K. M., and Cooper, H. (1998). The happy personality: a meta-analysis of 137 personality traits and subjective well-being. Psychol. Bull. 124, 197–229. doi: 10.1037/0033-2909.124.2.197

Diener, E., and Diener, M. (1995). Cross-cultural correlates of life satisfaction and self-esteem. J. Pers. Soc. Psychol. 68, 653–663. doi: 10.1037/0022-3514.68.4.653

Diener, E., Oishi, S., and Lucas, R. E. (2003). Personality, culture, and subjective well-being: emotional and cognitive evaluations of life. Ann. Rev. Psychol. 54, 403–425. doi: 10.1146/annurev.psych.54.101601.145056

Fei, X. T. (2002). Autonomy of resident: new target of community construction in urban China. Jianghai Acad. J. 3, 15–18.

Funder, D. C., and Fast, L. A. (2010). “Personality in social psychology,” in Handbook of Social Psychology , 5th Edn, eds S. T. Fiske, D. T. Gilbert, and G. Lindsey (New York, NY: Wiley), 668–697.

Gutiérrez, J. L. G., Jiménez, B. M., Hernández, E. G., and Pcn, C. (2005). Personality and subjective well-being: big five correlates and demographic variables. Pers. Individ. Differ. 38, 1561–1569. doi: 10.1016/j.paid.2004.09.015

Ha, S. E., and Kim, S. (2013). Personality and subjective well-being: evidence from South Korea. Soc. Indic. Res. 111, 341–359. doi: 10.1007/s11205-012-0009-9

Hartley, C., and Coffee, P. (2019). Perceived and received dimensional support: main and stress-buffering effects on dimensions of burnout. Front. Psychol. 10:1724. doi: 10.3389/fpsyg.2019.01724

Hayes, A. (2013). Introduction to Mediation, Moderation, and Conditional Process Analysis. New York, NY: Guilford.

Hayes, N., and Joseph, S. (2003). Big 5 correlates of three measures of subjective well-being. Pers. Individ. Differ. 34, 723–727. doi: 10.1016/s0191-8869(02)00057-0

Helliwell, J. F., and Putnam, R. D. (2004). The social context of well-being. Philos. Trans. Biol. Sci. 359, 1435–1446.

Hill, P. L., Turiano, N. A., Mroczek, D. K., and Roberts, B. W. (2012). Examining concurrent and longitudinal relations between personality traits and social well-being in adulthood. Soc. Psychol. Pers. Sci. 3, 698–705. doi: 10.1177/1948550611433888

Hofstede, G., and McCrae, R. R. (2004). Personality and culture revisited: linking traits and dimensions of culture. Cross Cult. Res. 38, 52–88. doi: 10.1177/1069397103259443

Inoue, Y., Funk, D. C., Wann, D. L., Yoshida, M., and Nakazawa, M. (2015). Team identification and postdisaster social well-being: the mediating role of social support. Group Dyn. Theory Res. Pract. 19:31. doi: 10.1037/gdn0000019

Jasinskaja-Lahti, I., Liebkind, K., Jaakkola, M., and Reuter, A. (2006). Perceived discrimination, social support networks, and psychological well-being among three immigrant groups. J. Cross Cult. Psychol. 37, 293–311. doi: 10.1177/0022022106286925

John, O. P., Donahue, E. M., and Kentle, R. L. (1991). The ‘Big Five’ Inventory: Versions 4a and 54. Technical Report. Berkeley, CA: Institute of Personality Assessment and Research.

John, O. P., and Srivastava, S. (1999). “The big five trait taxonomy: history, measurement, and theoretical perspectives,” in Handbook of Personality: Theory and Research , eds L. A. Pervin and O. P. John (New York, NY: Guilford Press), 102–138.

Joshanloo, M., Rastegar, P., and Bakhshi, A. (2012). The Big Five personality domains as predictors of social wellbeing in Iranian university students. J. Soc. Pers. Relat. 29, 639–660. doi: 10.1177/0265407512443432

Keyes, C. L. M. (1998). Social well-being. Soc. Psychol. Q. 61, 121–140.

Keyes, C. L. M. (2002). The mental health continuum: from languishing to flourishing in life. J. Health Soc. Behav. 43, 207–222. doi: 10.2307/3090197

Keyes, C. L. M. (2005). Mental health and/or mental illness? Investigating axioms of the complete state model of health. J. Consult. Clin. Psychol. 73, 539–548. doi: 10.1037/0022-006x.73.3.539

Keyes, C. L. M., and Ryff, C. D. (1998). “Generativity in adult lives: social structural contours and quality of life consequences,” in Generativity and Adult Development: Perspectives on Caring for and Contributing to the Next Generation , eds D. McAdams and E. de St. Aubin (Washington, DC: American Psychological Association), 227–263. doi: 10.1037/10288-007

Keyes, C. L. M., and Shapiro, A. D. (2004). “Social well-being in the United States: a descriptive epidemiology,” in How Healthy are We? A National Study of Well-Being at Midlife , eds O. G. Brim, C. D. Ryff, and R. C. Kessler (Chicago, IL: University of Chicago Press), 350–372.

Lee, R. M., Draper, M., and Lee, S. (2001). Social connectedness, dysfunctional interpersonal behaviors, and psychological distress: testing a mediator model. J. Counsel. Psychol. 48, 310–318. doi: 10.1037/0022-0167.48.3.310

Li, Q. M., and Chen, Z. X. (2015). Development level of big five personality in five dimensions and ten aspects: based on cross-sectional sample survey of 15–75 years old in China. Psychol. Sci. 1, 131–138.

Ma, X. (2013). The thinking way of Chinese and its transformation of modernity. J. Liaoning Administ. Coll. 15, 157–158.

Miao, Y. J., and Wang, Q. H. (2009). Investigation and Research on College Students’ social well-being. J. Gannan Norm. Univer. 30, 76–81.

Minkov, M., Fons, J. R., van de Vijver, O., and Michael, S. (2019). A test of a new short big-five tool in large probabilistic samples from 19 countries. Pers. Individ. Differ. 151:109519. doi: 10.1016/j.paid.2019.109519

Oh, H. J., Ozkaya, E., and LaRose, R. (2014). How does online social networking enhance life satisfaction? The relationships among online supportive interaction, affect, perceived social support, sense of community, and life satisfaction. Comput. Hum. Behav. 30, 69–78. doi: 10.1016/j.chb.2013.07.053

Podsakoff, P. M., MacKenzie, S. B., Lee, J. Y., and Podsakoff, N. P. (2003). Common method biases in behavioral research: a critical review of the literature and recommended remedies. J. Appl. Psychol. 88, 879–903. doi: 10.1037/0021-9010.88.5.879

Preacher, K. J., and Hayes, A. F. (2004). SPSS and SAS procedures for estimating indirect effects in simple mediation models. Behav. Res. Methods Instr. Comput. 36, 717–731. doi: 10.3758/bf03206553

Rhee, J., Seog, S. D., Bozorov, F., and Dedahanov, A. T. (2017). Organizational structure and employees’ innovative behavior: the mediating role of empowerment. Soc. Behav. Pers. Intern. J. 45, 1523–1536. doi: 10.2224/sbp.6433

Sarason, B. R., Sarason, I. G., and Pierce, G. R. (1990). “Traditional views of social support and their impact on assessment,” in Social Support: An Interactional View , eds B. R. Sarason, I. G. Sarason, and G. R. Pierce (New York, NY: Wiley), 9–25. doi: 10.1080/07349165.1982.9725924

Siedlecki, K. L., Salthouse, T. A., Oishi, S., and Jeswani, S. (2014). The relationship between social support and subjective well-being across age. Soc. Indic. Res. 117, 561–576. doi: 10.1007/s11205-013-0361-4

Sohi, K. K., Singh, P., and Bopanna, K. (2017). Ritual participation, sense of community, and social well-being: a study of seva in the sikh community. J. Relig. Health 57, 2066–2078. doi: 10.1007/s10943-017-0424-y

Sun, B. Y. (2016). Why is it possible for the urban community residents’ committees to “de-administrate”? Nanjing Soc. Sci. 7, 51–58.

Swickert, R. J., Hittner, J. B., and Foster, A. (2010). Big Five traits interact to predict perceived social support. Pers. Individ. Differ. 48, 736–741. doi: 10.1016/j.paid.2010.01.018

Taylor, S. E. (2011). “Social support: a review,” in Oxford Library of Psychology. The Oxford Handbook of Health Psychology , ed. H. S. Friedman (New York, NY: Oxford University Press), 189–214.

Udayar, S., Urbanaviciute, I., and Rossier, J. (2018). Perceived social support and Big Five personality traits in middle adulthood: a 4-year cross-lagged path analysis. Appl. Res. Q. Life 15, 395–414. doi: 10.1007/s11482-018-9694-0

Wong, S. T., Yoo, G. J., and Stewart, A. L. (2007). An empirical evaluation of social support and psychological well-being in older Chinese and Korean immigrants. Ethn. Health 12, 43–67. doi: 10.1080/13557850600824104

Zhang, W., Chen, Q., McCubbin, H., McCubbin, L., and Foley, S. (2011). Predictors of mental and physical health: individual and neighborhood levels of education, social well-being, and ethnicity. Health Place 17, 238–247. doi: 10.1016/j.healthplace.2010.10.008

Zhou, J. (2010). Structural validity analysis of Big Five Personality Inventory (BFI). Manag. Observ. 30, 252–253.

Keywords : big five personality, social support, social well-being, China, mediating effect

Citation: Yu Y, Zhao Y, Li D, Zhang J and Li J (2021) The Relationship Between Big Five Personality and Social Well-Being of Chinese Residents: The Mediating Effect of Social Support. Front. Psychol. 11:613659. doi: 10.3389/fpsyg.2020.613659

Received: 03 October 2020; Accepted: 31 December 2020; Published: 05 March 2021.

Reviewed by:

Copyright © 2021 Yu, Zhao, Li, Zhang and Li. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY) . The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

*Correspondence: Jiewei Li, [email protected]

† These authors share first authorship

Disclaimer: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.

• Open access
• Published: 10 July 2018

Predicting the Big Five personality traits from handwriting

• Mihai Gavrilescu   ORCID: orcid.org/0000-0002-4616-7362 1 &
• Nicolae Vizireanu 1  

EURASIP Journal on Image and Video Processing volume  2018 , Article number:  57 ( 2018 ) Cite this article

29k Accesses

44 Citations

13 Altmetric

Metrics details

We propose the first non-invasive three-layer architecture in literature based on neural networks that aims to determine the Big Five personality traits of an individual by analyzing offline handwriting. We also present the first database in literature that links the Big Five personality type with the handwriting features collected from 128 subjects containing both predefined and random texts. Testing our novel architecture on this database, we show that the predefined texts add more value if enforced on writers in the training stage, offering accuracies of 84.4% in intra-subject tests and 80.5% in inter-subject tests when the random dataset is used for testing purposes, up to 7% higher than when random datasets are used in the training phase. We obtain the highest prediction accuracy for Openness to Experience, Extraversion, and Neuroticism (over 84%), while for Conscientiousness and Agreeableness, the prediction accuracy is around 77%. Overall, our approach offers the highest accuracy compared with other state-of-the-art methods and results are computed in maximum 90 s, making the approach faster than the questionnaire or psychological interviews currently used for determining the Big Five personality traits. Our research also shows there are relationships between specific handwriting features and prediction with high accuracy of specific personality traits and this can be further exploited for improving, even more, the prediction accuracy of the proposed architecture.

1 Introduction

Handwriting has been used for centuries as a way of communication and expression for humans, but only recently its links to the brain activity and the psychological aspects of humans have been studied. The psychological study of handwriting with the purpose of determining the personality traits, psychological states, temperament, or the behavior of the writer is called graphology and is still a debatable domain as it lacks a standard, most of the handwriting interpretations being done subjectively by trained graphologists.

However, there have been various research papers showing the link between handwriting and neurological aspects of humans, one such study being the one of Plamondon [ 1 ], where it was shown that the brain forms characters based on habits of writers and each neurological brain pattern forms a distinctive neuromuscular movement which is similar for individuals with the same type of personality. Therefore, handwriting is, from this perspective, an accurate mirror of people’s brain.

Graphologists currently analyze multiple handwriting features in order to assess the psychological aspects of the writer, such as the weights of strokes [ 2 ], the trajectory of writing [ 3 ], the way the letter “t” or “y” are written [ 4 ], as well as other features related to how letters or words are written or how the text is positioned on the page.

In the current paper, we aim to build the first architecture in literature that is able to automatically analyze a set of handwriting features and evaluate the personality of the writer using the Five-Factor Model (FFM). To test this architecture, we propose the first database that links the FMM personality traits to handwriting features, which is a novel aspect of this research paper. The proposed system offers an attractive alternative to the standard FMM questionnaire or psychological interviews that are currently used for evaluating personality, because it is easier to use, it involves less effort, and is faster as well as removes the subjectivity from both subject’s (as usually the subject is asked to self-report on a specific questionnaire) as well as clinician’s sides (as typically psychologists are reviewing the questionnaire results and share opinions regarding the personality of the individual, opinions which can sometimes be prone to bias such that different psychologists might provide different evaluations). We show that our proposed system offers the highest accuracy compared to other state-of-the-art methods as well as share our findings regarding the relationship between several handwriting features and specific personality traits that can be further exploited to improve, even more, the accuracy of such a system.

In the following section, we present the state-of-the-art in the area of handwriting analysis, focusing on papers related to predicting the psychological traits of individuals. We continue in the subsequent section with describing the two models used (FMM and graphology analysis) followed by a detailed presentation of the three-layer architecture, as well as the classifiers and the structure of the neural network used. Finally, we detail the experimental results and share our findings and conclusions on the results obtained.

2 Related work

As mentioned previously, currently, there is no standard developed in predicting behavior based on handwriting, the majority of graphological analysis being done by specialized graphologists. However, research was conducted in the area of computer science which aimed to create such systems in order to recognize the behavior from handwriting in an easier way and also to standardize the graphological analysis. In the next paragraphs, we present the state-of-the-art in this area as well as several studies which made use of handwriting to determine the psychological traits or mental status of individuals.

Behnam Fallah and Hassan Khotanlou describe in [ 5 ] a research with a similar purpose as the one conducted in this paper, aiming to determine the personality of an individual by studying handwriting. The Minnesota Multiphasic Personality Inventory (MMPI) is used for training their system and a Hidden Markov Model (HMM) is employed for classifying the properties related to the target writer, while a neural network (NN) approach is used for classifying the properties which are not writer-related. The handwriting image is analyzed by these classifiers and compared with the patterns from the database, the output being provided in the form of the personality of the writer on the MMPI scale. Their system offers over 70% accuracy at this task. Similarly, in [ 4 ], an instrument for behavioral analysis is described with the task of predicting personality traits from handwriting. The approach takes into account the following handwriting features: letter “t,” lower loop of the letter “y,” the pen pressure, and the slant of writing. A rule-based classifier is used to assess the personality trait of the writer on the Myers-Briggs Type Indicator (MBTI) scale with also over 70% accuracy. The work of Chen and Tao [ 6 ] also provides an interesting exploratory study where they use combinations of Support Vector Machine (SVM), AdaBoost, and k-nearest neighbors (k-NN) classifiers for each of the seven personality dimensions in order to analyze a unique set of handwriting features. Their results are promising with accuracies ranging from 62.5 to 83.9%.

Although not aiming for personality traits, Siddiqi et al. [ 7 ] present a system that is able to predict the gender of individuals from scanned images containing their handwriting. A set of features is extracted from their writing samples, and artificial neural networks (ANNs) and Support Vector Machines (SVMs) are used to discriminate between the writing of a male and that of a female. The handwriting features employed are slant, curvature, texture, and legibility, computed in both local and global features. Evaluated on two databases under a number of scenarios, the system is able to predict with over 80% accuracy the gender of the writer. Similarly, in [ 8 ], it is proposed a way to describe handwritings based on geometric features which are combined using random forest algorithms and kernel discriminant analysis. The system is able to predict gender with 75.05%, age with 55.76%, and nationality with 53.66% when all the writers were asked to write the same text, and 73.59% for gender prediction, 60.62% for age prediction, and 47.98% for nationality prediction when each subject wrote a different text.

Another interesting research is the one conducted by Gil Luria and Sara Rosenblum [ 9 ] which uses handwriting behavior in order to determine the characteristics of both low and high mental workloads. They asked 56 participants to write three arithmetic progressions of different difficulties on a digitizer, and differences are seen in temporal, spatial as well as angular velocity spaces, but less in the pressure space. Using data reduction, they identify three clusters of handwriting types and conclude that handwriting behavior is affected by the mental workload. Zaarour et al. [ 10 ] show another interesting research where handwriting is employed to improve the performance of pupils through a system which takes as input different drawings and writings and, by means of a Bayesian network-based model, they can determine the writing style of the child which can be further analyzed by a child psychologist in order to advise parents on how to improve their child education. Similarly, Sudirman et al. [ 11 ] present a system that studies the behavior of children based on their handwriting, starting from the assumption that children are the best subjects to be analyzed in the context of handwriting as they are less influenced by cultural background and their cognition rate is evolving very fast. Therefore an automatic system is built which aims to determine the developmental disorders that the children might be suffering from, with accuracies of over 78%, making the approach attractive for both teachers as well as therapists for patients’ monitoring. Researchers in [ 12 ] present a system tasked with decreasing the time for job candidate selection in the pre-employment stage using automatic personality screening based on visual, audio, and lexical cues. The system extracts a set of relevant features which are used by a chain of machine learning techniques in order to predict candidates’ scores on the Five-Factor Model scale and a classifier is used to combine the prediction results from all the three cues. The experimental results show promising results in terms of performance on first impression database.

Another direction for many studies involving handwriting analysis is the detection of deceit. Luria et al. [ 13 ] show such research where a non-intrusive system analyzes the handwriting in the context of healthcare with the purpose of detecting the false information that patients provide about their health. As current ways of determining deception are invasive and do not comply with a clinician-patient relationship, such an approach of using the handwriting as a tool is attractive from research perspectives. Subjects participating in the experiment were asked to write true/false statements about their medical condition on a paper linked to a digitizer. After this first step, the deceptive and truthful writings of all the subjects are compared and used to divide the subjects into three groups according to their handwriting profiles. It is found that the deceptive writing takes longer to write and is broader and the two types of writings show significant differences in both spatial and temporal vectors. In [ 14 ], similar research is conducted, based on the same assumption that for people it is easier to tell the truth than to lie; hence, we need to see changes in both velocity and temporal spaces when analyzing the handwriting features. Conducted in 11 languages, this research demonstrates the same point as in [ 13 ], with the specific purpose of helping managers pinpoint sudden emotional changes and decode handwritten messages to reveal the true meaning of those messages as well as detect lies.

Besides detecting deceit, the handwriting is also used for predicting physical diseases. Researchers in [ 15 ] present a study where diabetics’ disease can be predicted with over 80% accuracy from handwriting. Similarly, in [ 16 ], the handwriting is used to predict micrographia (the decrease in the size of letters as well as the velocity and acceleration of writing) that is commonly associated with Parkinson’s disease (PD). The system, tested on PD-diagnosed patients, offers over 80% accuracy on 75 tested subjects. The study described in [ 17 ] is another research analyzing the link between the handwriting and children with autism spectrum disorder (ASD), knowing the fact that children with ASD have several weaknesses in handwriting. Boys aged 8–12 years and diagnosed with ASD were asked to take a digitized task in order to determine the handwriting performance using advanced descriptive methods. The study shows moderate to large links between handwriting performance and attention, ASD symptoms and motor proficiency, providing a relationship between handwriting and the ASD symptoms in terms of severity, attention, and motor behaviors.

Since handwriting analysis is a complex task requiring multiple techniques in order to analyze the multitude of handwriting features, there is a wide range of methods typically employed. For offline handwriting analysis, the normalization of the handwritten sample is the first step in order to ensure any possible noise is filtered out. As part of normalization phase, methods for removing the background noise (morphological approaches or Boolean filters are typically used [ 18 ]), sharpening (Laplace filters, Gradient masking or unsharp masking [ 19 ]), and contrast enhancement (unsharp mask filters [ 20 ]) are essential for ensuring the analysis of the handwriting is done with high accuracy. Also, as the contour of the written letters is essential for this task, methods for contour smoothing also need to be used, the most common ones being the local weighted averaging methods [ 21 ]. After all these processing steps are applied to the handwritten sample, the image needs to be compressed and converted to greyscale and different types of thresholding techniques can be employed for this step [ 22 ]. Post-compression, the written text needs to be delimited through page segmentation methods where techniques for examining the foreground and background regions are employed, the most common one being the white space rectangles segmentation [ 23 ]. One of the most challenging tasks is the one of segmenting the handwritten image into text lines and words. For this, the Vertical Projection Profile [ 24 ] method has shown the most promising results and this is the one that we use in this paper for both row and word segmentation. Regarding feature classification, different classifiers are used successfully for each of the handwriting features. For example, for lowercase letters “t” and “f,” the most common method used is template matching, for writing pressure gray-level thresholding methods are employed [ 22 ], while for connecting strokes the Stroke Width Transform (SWT) has shown the best classification accuracy compared to other state-of-the-art methods. In the following sections, we present in detail the classifiers used for each of the handwriting features analyzed in the current paper.

With all these in mind, the current research proposes a novel non-invasive neural network-based architecture for predicting the Big Five personality traits of a subject by only analyzing handwriting. This system would serve as an attractive alternative to the extensive questionnaire typically used to assess the FMM personality traits and which is usually cumbersome and non-practical, as well as avoid the use of invasive sensors. We focus our attention on handwriting because it is an activity familiar to almost everyone and can be acquired fast and often.

In the next section, we present the theoretical model and the architecture of our system.

3.1 Theoretical model

As mentioned in the previous section, our research is proposing a novel non-invasive neural network-based architecture for predicting the Big Five personality traits of an individual solely based on handwriting. Therefore, our study is based on two psychological tools: Big Five (Five-Factor Model—FMM) [ 17 ] and graphological analysis. We detail both these instruments in the next subsections.

3.1.1 Big Five (Five-Factor Model)

Big Five (Five-Factor Model) [ 25 ] is a well-known model for describing the personality of an individual. It is based on five basic personality traits which are grouped in sub-factors, as follows:

Openness to Experience : refers to people who can easily express their emotions and have a desire for adventure, appreciation for art, and out-of-the-box ideas. Typically, on this scale, people are rated based on the dichotomy: consistent vs. curious ;

Conscientiousness : refers to people who are dependable, have a predilection towards behaviors which are carefully planned, and are oriented towards results and achievements. On this scale, people are rated based on the dichotomy: organized vs. careless ;

Extraversion : refers to people who easily express positive emotions, like other’s people company, are assertive, and talkative. On this scale, people are rated on the dichotomy: outgoing vs. solitary ;

Agreeableness : refers to people who have a tendency to be compassionate instead of suspicious, as well as helpful, and tempered. On this scale, people are rated based on the dichotomy: compassionate vs. detached ;

Neuroticism: refers to people who lack emotional stability and control and tend to experience negative emotions easily, such as anger and anxiety, as well as a vulnerability to depression. On this scale, people are rated based on the dichotomy: nervous vs. confident.

FMM is successfully used on a wide variety of tasks. The research conducted in [ 26 ] shows that compared to other methods for assessing the personality of an individual, FMM offers more stability over time, the Big Five personality types reaching their stability peak 4 years after starting work. FMM has also proved to be useful in determining personality disorders, such as depression or anxiety, and even substance use, and was shown to be an indicator for different physical diseases, such as heart problems, cancer, diabetes or respiratory issues [ 27 ]. It is also successfully used in the area of career development and counseling as well as team performance, but also for improving learning styles and the academic performances of students [ 28 ]. Because of its extensive use and broad perspective of applications we employ it in our current study.

3.1.2 Graphological analysis

Typically, when analyzing the handwriting of an individual, graphologists are looking for a specific set of features, each of them conveying a specific message [ 29 ]. The main handwriting features used and the ones that we explore in the current paper are the following: baseline, word slant, writing pressure, connecting strokes, space between lines, lowercase letter “t,” and lowercase letter “f.” Examples of each of these features and their types as explained in [ 30 ] can be observed in Table  1 .

The baseline of the handwriting refers to the line on which the written words flow. It is further divided into ascending baseline (associated with optimistic people), descending baseline (associated with pessimistic people and over-thinkers), and leveled (associated with people with high levels of self-control and reasoning).

The word slant refers to how the words are written in terms of inclination/slant. Possible slant types are the following: vertical slant (associated with people who can easily control their emotions), moderate left slant (associated with people who find it hard to express emotions), extreme left slant (associated with people who want to be in permanent control and suffer from self-rejection), moderate right slant (associated with people who can easily exteriorize their emotions and opinions), and extreme right slant (associated with people who are impulsive and lack self-control).

The writing pressure refers to the amount of pressure that is applied to the pen on the paper: light writer (refers to people who hardly get affected by traumas), medium writer (refers to people who are usually affected by pain or traumas), and heavy writer (refers to people who are deeply affected by traumas and emotions).

Connecting strokes refer to how the letters composing words are connected to each other. These are dichotomized into not connected (refers to people that can hardly adapt to change), medium connectivity (refers to people who can adapt to change as well as like changing environments), and connected letters (refers to people who can quickly adapt to change).

Lowercase letter “t” typically refers to how the t-bar on the letter “t” is written. If it is written very low , it is an indication of low self-esteem, if it is written very high it is an indicator of high self-esteem.

Lowercase letter “f” refers to how the letter “f” is written. If it has an angular point , the person can be easily revolted, if it has an angular loop , the person has a strong reaction to obstacles, if it has a narrow upper loop it is usually associated with narrow-minded people, if it is cross-like it is associated with an increased level of concentration, and if it is balanced it is an indicator of leadership abilities.

Spaces between lines refer to the space left by the writer between two consecutive lines. We can have lines separated, evenly spaced (associated with people who can organize work and have clear thoughts) or lines crowded together with overlapping loops (associated with people with confused thinking and poor organizational skills).

3.2 Proposed architecture

We design the architecture on three layers as follows: a base layer where the handwriting sample is normalized and the handwriting features are acquired, an intermediary layer where a Handwriting Map is built based on the handwriting features provided by the base layer, and a top layer where a neural network is used in order to determine the Big Five personality type of the writer. In the following subsections, we present each of these layers in detail.

3.2.1 Base layer

The base layer has the primary purpose of converting the scanned handwriting in the set of handwriting features mentioned in previous sections. A flowchart of the central processing blocks of this layer can be observed in Fig.  1 .

Flowchart of the base layer and handwriting features extraction

The main steps are detailed below:

Normalization:

◦ Noise reduction : in order to remove the noise added by the scanning device or the writing instrument which typically cause distortion, disconnected strokes or unwanted lines or points, we use three filters. Boolean filters are used for removing the textured background as they were shown to outperform other morphological methods for cases when the text is written on highly texturized backgrounds both in terms of accuracy and processing time [ 18 ]. For sharpening, we use the ramp width reduction filter as it is known as the most effective algorithm for ramp edge sharpening [ 19 ]. Adaptive unsharp masking is employed for adjusting the contrast [ 20 ] which is widely used as an effective method for contrast enhancement.

◦ Contour smoothing : in order to reduce the possible errors that appear due to unwanted movement of writer’s hand during writing we use an optimal local weighted averaging method [ 21 ] ensuring that these glitches are filtered out and only the strokes relevant for our analysis are kept. We opted for this algorithm as opposed to other less complex local weighted averaging methods because this method is known to provide more accurate estimations of contour point positions, tangent slopes, or deviation angles which are essential for our handwriting analysis task.

◦ Compression : we used global thresholding in order to convert the color images to binary. We used the histogram modified by integral ratio [ 22 ] in order to determine the global threshold value as it was shown to provide better performance compared to other compression techniques.

◦ Isolation of handwriting in the page : in order to only keep the handwritten text for the next steps of our handwriting analysis task, we use the white space thinning method [ 23 ] as it is a simple and fast method for this task; hence, we cut the page recursively on the two dimensions until only the handwritten text is delimited.

Row segmentation : For row segmentation, we use the Vertical Projection Profile (VPP) method [ 24 ] as it was showed to provide the best classification accuracy compared to other row and word segmentation methods. We, therefore, analyze the sum of pixels for each row in the image and determine as row boundaries those with a sum lower than 8% of the highest pixel sum in the text sample. The threshold of 8% was chosen through trial-and-error after conducting tests on 100 handwriting samples using a leave-one-out approach and the average accuracy for correct row segmentation was 98.5%. Following this step, every row in the handwritten text has a corresponding bounding rectangle.

◦ Spacing between lines feature : based on the bounding rectangles delimiting each row from handwriting, we determine the amount of overlap between two consecutive rows. If the overlap is higher than 15% of the sum of both row bounding rectangles’ surfaces, we consider that the rows are crowded together , otherwise, they are considered evenly spaced . The 15% threshold was determined to be optimal for ensuring over 98% accurate classification of this handwriting feature.

◦ Baseline feature : in order to determine the baseline features for each row, we use the method depicted in [ 31 ] where we study the pixel density of each segmented row rectangle and we rotate the rectangle within the − 30° and + 30° angle thresholds until the highest pixel density is horizontally centered. This method is broadly used for baseline feature extraction offering higher classification accuracy and faster convergence compared to other state-of-the-art methods. If the rotation needed to align the highest pixel density horizontally is within [− 5°; + 5°], we consider that we have a leveled baseline , if it is within [− 30°; − 5°], an ascending baseline, and within [+ 5°; + 30°] a descending baseline .

◦ Writing pressure feature : we use the standard gray-level thresholding method that is widely used for the task of writing pressure classification [ 32 ] with high accuracy and fast convergence. We analyze the grayscale values for the segmented rectangle containing the row and we calculate the average for the segmented row. The result is classified as light writer for a value within 25 and 50%, medium writer for a value within 10 and 25%, and heavy writer for a value within 0% (absolute black) and 10%.

Word segmentation : In order to further segment the words in a row, we use the same VPP method [ 24 ] that we employed for row segmentation as it was shown to provide better classification results than other state-of-the-art methods. We compute the height of the row first and use it for comparison purposes in order to determine whether a space between two strokes is indeed an inter-word space or not. We generate a vertical projection profile where we determine the pixel density for each vertical column and we determine the columns with low density, which are considered candidates for spaces between words. As there are cases when such gaps might not correspond to actual word separation spaces, we consider them spaces only if the number of consecutive columns with low density is not lower than 10% of the row height. The 10% threshold was determined through trial-and-error after testing the algorithm on 100 handwritten samples and obtaining the highest word segmentation accuracy of 98.2%. The segmented words are bounded by rectangles similarly as in the row segmentation case.

◦ Word slant feature : in order to determine the word slant feature, we use the same technique described in [ 33 ]. We calculate the vertical pixel density histogram for each angle within [− 20°; + 20°] and for each column in the histogram we determine the number of pixels and divide it with the highest and lowest pixel in the analyzed word segment. The values from all columns are then summed and the angle where the computed sum is the highest is considered to be the slant of the writing. We then classify the word slant as follows: if the angle is within [− 2.5°; + 2.5°], it is a vertical slant ; if it is within [− 7.5°; − 2.5°], it is a moderate left slant ; if it is lower than − 7.5°, it is an extreme left slant ; if it is within [+ 2.5°; + 7.5°], it is a moderate right slant ; and if it is higher than + 7.5°, an extreme right slant .

Letter segmentation : for segmenting the letters from each delimited word segment, we use the stroke width transform (SWT) [ 34 ] method for determining the average stroke width of the word. We use this operator because it is local and data dependent, making it faster and more robust than other methods that need multi-scale computations. We then create a projected profile for the word segment and determine the columns where the projection value is lower than 8% than the highest projected value in the word. For the identified strokes, we determine their width and compare it with the word’s average stroke width. If it is lower than 50%, we create a bounding box surrounding the character and we crop out the bounding box from the word segment. The 50% threshold was determined to be optimal after testing the method on 100 handwritten samples and obtaining 98.2% accuracy for letter segmentation. With the remaining part of the word segment, the process is repeated until all letters are identified.

◦ Connecting strokes feature : in order to compute the connecting strokes feature, we use the letter segmentation algorithm previously described and we compare each stroke width connecting two consecutive letter bounding boxes with the average stroke width of the word. If the stroke width is below 10% of the average stroke width of the word, we consider it as not connected ; if it is above 30%, we consider it connected ; and if it is between 10 and 30%, it is considered as having medium connectivity.

◦ Lowercase letter “t” feature : as letters are now delimited in corresponding bounding boxes, we use template matching to compare each letter to a set of predefined templates of letter “t” from the Modified National Institute of Standards and Technology (MNIST) database [ 35 ]. The templates were previously divided into the two categories of letter “t” ( very low “t” bar and very high “t” bar ), and we use Euclidean similarity to measure the letter matching to the chosen MNIST prototypes. The threshold matching determined as optimal through trial-and-error is 0.88 and the accuracy for detecting the right letter “t,” tested on 100 handwriting samples with a leave-one-out approach, is 98.2%.

◦ Lowercase letter “f” feature : we use the same method depicted for letter “t” with the difference that the letter “f” templates from the MNIST database are divided into five categories corresponding to the ones analyzed (angular point, angular loop, narrow upper loop, cross-like and balanced). The threshold, in this case, is 0.92 corresponding to an accuracy of 97.5%.

3.2.2 Intermediary layer (Handwriting Map)

As we previously mentioned, the base layer offers as inputs to the intermediary layer the handwriting feature types for each letter in the exemplar. These are coded in the Handwriting Map (HM) using a binary code. Therefore if, for example, connecting strokes have medium connectivity, the code for this is 010 (0—connected, 1—medium connectivity, 0—not strongly connected). Typically, for each analyzed letter, we have the following possible codes associated with each of the seven handwriting features that all compose one row in the HM:

Baseline : position 1 to 3 : possible values are 100—ascending, 010—descending, 001—leveled;

Connecting strokes : position 4 to 6 ; possible values are 100—not connected, 010—medium connectivity, 001—strongly connected;

Word slant : position 7 to 11 ; possible values are 10000—vertical slant, 01000—moderate left slant, 00100—extreme left slant, 00010—moderate right slant, 00001—extreme right slant;

Writing pressure : position 12 to 14 ; possible values are 100—light writer, 010—medium writer, 001—heavy writer;

Lowercase letter “t” : position 15 to 16 ; possible values are 10—very high; 01—very low; 00—not a lowercase letter “t”;

Lowercase letter “f” : position 17 to 21 ; possible values are 10000—cross-like, 01000—angular loop, 00100—angular point, 00010—narrow upper loop, 00001—balanced; 00000—not a lowercase letter “f”;

Space between the lines : position 22 to 23 ; possible values are 10—evenly spaced, 01—crowded together.

Therefore any row entry in the map has the following structure: [100][010][00010][100][00][00010][10] (which means ascending baseline —100, medium strokes connectivity —010, moderate right slant —00010, light writer —100, not a lowercase letter “t” —00, Narrow Upper Loop on lowercase letter “f” —00010, evenly spaced lines —10).

Two observations should be made about the above-constructed mapping:

For baseline , we might have the same code for all letters;

For space between the lines , we might have the same code for all letters that are associated with a row in the handwritten sample.

Therefore, each letter in the handwriting sample generates a row in the HM in the form of a binary code which is then used in the top layer in a pattern recognition task in order to determine the Big Five personality traits.

3.2.3 Top layer

As we have detailed earlier, we have an HM that contains for each letter its handwriting features in the form of a binary code. Therefore, the HM is a matrix containing all the letters in the handwriting exemplar together with their coded features and based on this the system should be able to determine the Big Five personality trait of the writer.

As the task is a pattern recognition task and also considering that our architecture is bottom-up with no feedback loops, we use a feed-forward neural network. Also, with the same premises in mind, the training method used is backpropagation , which has proven to be very effective and offers fast learning in similar cases [ 36 ].

We define only one neural network that is called the Five-Factor Model–Neural Network (FFM-NN). In order to avoid overfitting it by fetching all the letters from the exemplar, we fetch them by rows and we consider that we do not have more than 70 letters on each row. If a row in the handwritten sample has more than 70 letters, only the first 70 are analyzed. More than this, this approach offers the ability to have multiple tests done on the neural network and we can average the results in order to reach more conclusive ones. As we have 23 entries for each row in the HM, in total we have 1610 input nodes in FFM-NN.

The output layer contains five nodes for each of the five dimensions of FMM. Each node computes a 0 if the subject is found on the lower side of the analyzed dimension, and 1 if it is found on the higher side of the dimension (e.g., a 1 for Openness to Experience means that the subject is more curious than consistent, while a 0 for Neuroticism means that the subject is more inclined towards being nervous than confident).

If we consider N in the number of input training vectors and an N-dimensional set of input vectors for the FFM-NN neural network X FFM  −  NN  = { x FFM  −  NN n }, n  = 1, 2… N in , so that x FFM  −  NN  = [ x FFM  −  NN 1 ,  x FFM  −  NN 2 … x FFM  −  NN N ] T , and a K out the number of output vectors and K-dimensional set of output vectors Y FMM  −  NN  = { y FMM  −  NN k }, k  = 1, 2… K out so that y FFM  −  NN  = [ y FFM  −  NN 1 ,  y FFM  −  NN 2 , …,  y FFM  −  NN K ] T , and if we denote the matrix of weights between input and hidden nodes, W FFM  −  NNH , the matrix of weights between the hidden nodes and the output nodes W FFM  −  NNO with L the number of hidden nodes, and f FMM  −  NN 1 a   and f FMM  −  NN 2 a the activation functions, the expression form for the output vectors can be written as follows:

The input features for each letter on a row is fetched to the input nodes which then distributes the information to the hidden nodes and computes the weighted sum of inputs sending the result to the output layer through the activation function. In backpropagation stage, the Average Absolute Relative Error (AARE) (2) is calculated as the difference between what is expected ( y FMM  −  NN e ) and what is determined ( y FMM  −  NN p   with p  = 1, 2… N in ) and W FMM  −  NNH and W FMM  −  NNO weight matrices are calibrated in order to minimize the AARE FMM  −  NN :

With the purpose of +/− balance in the hidden layer, the activation function chosen for the input layer is tanh, also considering it offers fast convergence and has a stronger gradient than the sigmoid function. Because the final task of the neural network is a predictive one, we use sigmoid as activation function for the hidden layer, taking into account its non-linearity and that its output is in the range of [0,1]. Conducting various tests, through trial-and-error, we determined that the optimal number of hidden nodes in order to avoid overfitting is 1850. The optimal learning rate is determined as 0.02, the optimal momentum is 0.4, and 200,000 training epochs are needed to train the system in an average of 8 h on an Intel i7 processor computer. We use Gradient Descent to learn the weights and biases of the neural network until AARE is minimized and, in order to ensure an even spread of the initial weights, we use the Nguyen-Widrow weight initialization. The structure of the neural network can be observed in Fig.  2 .

FMM—neural network structure

3.3 Overall architecture

3.3.1 training database and handwriting text samples.

For testing the above-described architecture, we create our database containing both handwritten exemplars as well as the FMM personality trait of the writer. In collecting this, we involved 128 individuals, out of which 64 were males and 64 females, with ages between 18 and 35, all of them participating to this experiment in accordance and aware of the Helsinki Ethical Declaration.

Each of the 128 subjects was asked to take the FMM questionnaire as well as provide six handwriting samples. The FMM questionnaire results were analyzed by specialized psychologists to assess their results on the five personality dimensions. In what it concerns the six handwriting samples, two of them are a predefined text representing the London Letter [ 32 ], a standard exemplar broadly used by graphologists for handwriting analysis, while the others are minimum 300 words texts that subjects could write freely and randomly. All text samples are collected in the English language.

To summarize, for each subject involved in training we have their corresponding FMM personality dimensions results as well as six handwriting samples, out of which two are the London Letter.

In Fig.  3 we can observe an example of the London Letter collected from one of the subjects. The London Letter is chosen because of the handwriting features that we are collecting, such that lowercase letter “t” is assessed at the beginning (e.g., “ t o”, “ t hen”, “ t onight”), middle (e.g., “Swi t zerland”, “Le tt ers”), and end (e.g., “quie t ”, “expec t ”) of words, lowercase letter “f” is analyzed at the beginning of words (e.g., “ f or”) or intercalated (e.g., “le f t”) as well as other situations that pose difficulties to writers and help us better discriminate between other handwriting features, such as: words starting with uppercase (e.g., Zermott Street), group of longer words (e.g., “Athens, Greece, November”), words containing doubled letters (e.g., “Greece”, “Zermott”), use of letters that need additional strokes (such as x, z, i, j; e.g., “E x press”, “Swit z erland”, “V i enna”, “ j o i n”), and intercalating numbers and/or punctuation (e.g., “King James Blv. 3580.”).

Handwritten sample of The London Letter

In the following section, we present the training as well as testing stages and how they use the above-described database.

3.3.2 Training and testing phases

The proposed architecture is built using 55,000 code lines in Scala programming with Spark Library. The testbed is functioning on an i7 processor with 8GB of RAM and it is designed to work in two stages: training and testing. The overall architecture can be seen in Fig.  4 .

Proposed architecture—overview

In the training stage , the FMM-NN needs to be trained to learn the handwriting patterns and compute the right values for the five personality dimensions. We, therefore, use a set of handwriting samples as training samples that are fetched to the base layer. The handwriting samples are first normalized, then the words are split into letters and the handwriting features for each letter are extracted and sent to the intermediary layer. In the intermediary layer, the HM is built which contains a row for each letter from the handwritten sample in the form of binary codes as previously presented. Every time we have handwritten features collected for 70 new letters, these are fetched to the FMM-NN which is trained via backpropagation so that its output is the one obtained from the FMM questionnaire. When AARE is low enough and the training samples are finished the system is considered trained and can be tested.

In the testing stage , the analyzed handwriting exemplar is also normalized and split into letters in the base layer. The letters are then analyzed and their features are determined and sent to the intermediary layer which computes the HM. When 70 new letters are computed in HM, these are sent to the FMM-NN which provides an output representing its predicted FMM personality dimensions in the form of five binary codes, as previously explained. When on five consecutive rows (five sets of 70 letters) we have the same binary codes, the system considers that those are the personality dimensions of the writer and outputs the final result. If there are no five consecutive rows generating the same binary output (meaning that different personality traits are detected in any five consecutive rows), the result is flagged as Undefined. We chose five consecutive rows as they correspond to an average sized word (of five letters) and we determined that reducing or increasing this threshold results in lower system accuracies.

In the next section, we show the experimental results after testing the architecture as well as a comparison with state of the art.

4 Experimental results and discussion

As we described previously, due to the lack of a publicly available database that would relate the handwriting features with FMM, we built our database to support this study. The database contains handwritings collected from 128 subjects (64 females and 64 males), with ages between 18 and 35 years old as well as their results after filling in the FMM questionnaire which was subsequently analyzed by specialized psychologists to ensure the FMM personality traits are evaluated correctly. For testing the degree of generalization of the proposed approach when dealing with random handwritten text and the influence of the predefined handwritten text in both training and test phases, the database is divided into two main datasets: controlled dataset (consisting of handwriting samples where subjects were asked to write a predefined text—the London letter), and the random dataset (consisting of handwriting samples where subjects wrote a minimum 300 words text freely). Also for testing purposes, in order to determine the ability of the proposed approach to recognize the FMM features of a writer that was not involved in training, we divide the database in writer-specific datasets which contain handwritings only from one specific writer. Each sample from the database is therefore tagged with both the type of dataset to which it pertains (controlled or random) as well as a unique code specifying the writer. The tests conducted in both the intra-subject and inter-subject methodologies are presented in the following sections.

4.1 Own database tests

4.1.1 intra-subject methodology.

Intra-subject methodology refers to training and testing the system on handwriting samples coming from the same writer. We, therefore, use n -fold cross-validation for each writer-specific dataset taking also into account the handwriting type (controlled or random). For example, for determining the accuracy of the method when the controlled dataset is used both in test and training phases, since we have only two samples for each writer, we use leave-one-out cross-validation where one of the samples is involved in training and the other is used for testing and vice-versa. Similarly, for determining the accuracy of the method when the random dataset is used for training and the controlled dataset of testing, we train the system on the writer-specific random dataset (containing four samples) and we test it on the writer-specific controlled dataset (containing two samples) via n -fold cross-validation. The tests are repeated for all 128 users and the results are averaged and are detailed in Table  2 .

We observe the highest prediction accuracy when the system is trained and tested on the controlled dataset reaching 85.3% prediction accuracy, however when we use the same controlled dataset for training and we test the proposed approach on samples from the random dataset the accuracy does not decrease by much, reaching 84.4%. This is an important observation as it shows that the need for predefined handwritten texts is only for training purposes, while for testing we can use random texts which perform roughly similar to the predefined one. Similarly, when the controlled dataset is used for training, the cases where the personality type is flagged as Undefined is the lowest (0.2%), also sustaining the idea that the controlled dataset adds more value to the prediction accuracy when used in training stage as opposed to the random one. This indicates that if the text exemplar used for training handwriting samples is adequately chosen in order to train the neural network on all the analyzed features, using such an application we do not need a standard text for testing and we can ask the subject to write any text they like, making the approach more flexible and easy to use.

The highest prediction accuracies are obtained for Openness to Experience (88.3%—when the system was trained on the controlled dataset and tested on the random dataset), followed by Extraversion (87.4%), Neuroticism (85.3%), while for Conscientiousness and Agreeableness the results are lower, around 80%.

The average number of rows needed to compute the FMM personality types is 9 for the case where the controlled dataset is used in training and the random one for testing and maximum 14 when the random dataset is used for training. Typically, for a row to be computed it takes an average of 5 s, hence the system provides the FMM personality type in no more than 45 s when the controlled dataset is used in training, making the approach fast and attractive for clinicians as an alternative to the FMM questionnaire or psychological interviews.

4.1.2 Inter-subject methodology

In inter-subject methodology, we train the system with handwriting samples coming from different writers than those used for testing in order to determine the ability of the proposed approach to extrapolate the trained data to new writers. We used n -fold cross-validation, keeping the database division in controlled and random datasets, and ensuring that handwritings from the writer tested have not been used for training. For example, for training the accuracy of the system when trained on handwritings containing a predefined text (controlled dataset) and testing on handwritings with random text in inter-subject methodology, we use the controlled handwritings from all subjects except the one used for training (2 controlled samples/subject × 127 subjects used in test = 254 samples), and we test using n -fold cross-validation on random handwriting samples from the remaining subject (four samples). The tests are repeated until all subjects and all their samples are used in the testing phase and the averaged results are detailed in Table  3 . To note that we also conduct several tests where we decreased the number of subjects involved in training in order to analyze the change in accuracy when the number of subjects is increased.

Similarly to the intra-subject methodology, the highest prediction accuracy is obtained when the controlled dataset is used for both training and testing and when the system is trained on the highest number of subjects. In this case, the overall prediction accuracy is 84.5%. It is interesting to observe that reducing the number of subjects involved in training does not result in many decreases in terms of prediction accuracy, such that if we use only 96 subjects in training the prediction accuracy is 1.8% lower and when we use 64 subjects in training it decreases with about 1.6% more. This little decrease as well as the fact that high accuracies are obtained when the controlled dataset is used for training and the random dataset for test (78.6%) compared to when the random datasets are used for both training and testing (when the prediction accuracy was 6% lower) provides the same conclusion as in the intra-subject methodology, that the controlled dataset adds more value to the performance of the system if used in the training stage, helping the system learn better the handwriting features. Once learned, for testing purposes random texts can be used in the handwriting sample, providing only 5% lower accuracy, but making the system more practical (in the sense that the subject can write freely whatever text he/she wants). Similarly as in the case for intra-subject methodology, the fact that the number of cases where the personality type is flagged as Undefined is lower when the controlled dataset is used for training, with a maximum of just 0.7%, is another indicator that using the controlled dataset in the training stage improves the prediction accuracy by improving system’s ability to discriminate between different FMM personality types.

As in the intra-subject tests, in inter-subject ones, the highest prediction accuracy is obtained for Openness to Experience (88.6%), Extraversion (87.1%), and Neuroticism (86.3%), while lower accuracies are obtained for Consciousness and Agreeableness, roughly around 80%. When controlled datasets are used for training, the average number of rows needed to determine the personality types is 12 taking around 60 s which supports the idea that the proposed approach is fast and can be an attractive alternative to the FMM questionnaire or psychological interviews commonly used for evaluating the FMM personality types.

4.1.3 Relationship between the handwriting features and FMM

We conduct the next experiment in order to see which handwriting feature is associated with each of the five personality traits in FMM. In order to accomplish this, we create a background application that checks the HM and counts each occurrence of all the handwriting feature classifications against each of the five personality traits. This is acquired with the system trained on controlled datasets for 127 subjects and tested on the random datasets for the remaining subject with n -fold cross-validation, averaging the results. The results obtained are highlighted in Table  4 .

It can be observed that there are several links between the five personality types and the handwriting features, such that extreme left word slant, descending baseline, and cross-like lowercase letter “f” are associated with Conscientiousness, while medium connected strokes, medium right word slant, and balanced lowercase letter “f” are associated with Openness to Experience. These findings are significant as they can be used to optimize the proposed architecture such that the neural network is trained and tested only on the handwriting features that have relevant information about the personality traits that are investigated, the others being filtered out.

4.2 Comparison with state-of-the-art

As currently there is no standard public database that is broadly used for testing and comparing different architectures and methods for evaluating personality evaluation based on handwriting, we test the most common methods for assessing personality from handwriting on our database and compare the results with those obtained from our proposed approach. As it can be observed, our approach offers 84.4% accuracy for intra-subject tests and 80.5% accuracy for inter-subject tests, surpassing the rule-based classifier approach of Champa and AnandaKumar [ 4 ] with 12.5%, as well as the SVM, k-NN, and Ada-Boost combination of classifiers employed by Chen and Lin in [ 6 ], with 7.2%, respectively. Similarly, the proposed approach performs slightly better at the task of determining the FMM personality traits based on handwriting compared with the HMM-NN combination employed by Fallah and Khotanlou [ 5 ]. The results are detailed in Table  5 .

5 Conclusions

We described the first non-invasive three-layer architecture in literature that aims to determine the Big Five personality type of individuals solely by analyzing their handwriting. This novel architecture has a base layer where the handwritten sample in the form of a scanned image is normalized, segmented in rows, words, and letters and based on the computed segments the handwriting features are determined; an intermediary layer where a Handwriting Map (HM) is computed by binary coding the handwriting feature type of each letter; and a top layer where a feed-forward neural network is trained via backpropagation to learn the patterns from the HM map and compute the FMM personality traits.

In order to train and test this novel architecture, due to lack of any database that would link the FMM personality traits with handwriting samples, we create the first such database containing the FMM personality traits of 128 subjects and six handwriting samples from each of them with both predefined text (referred to as controlled dataset) as well as random text freely chosen by subjects (referred to as random dataset). We test our novel architecture on this database in both intra-subject and inter-subject methodologies and we obtain the highest prediction accuracies when the controlled dataset is used in the training stage, which shows that choosing a predefined text to be used for training the system is an important point in order to reach high accuracies, while testing can be done on random texts with no essential need for predefined texts to be used. This is an essential finding for real applications of such a systems, as it provides flexibility to the end-user, such that he/she will not have to write a predefined text every time, instead writing it only at the beginning in order to train the system, and then, to evaluate his/her personality traits at any given moment of time, he/she can use any random text he/she wants. In intra-subject tests, when the controlled dataset is used for training and random dataset for testing, we obtain an overall accuracy of 84.4%, while in inter-subject tests with a similar test-case we obtain an overall prediction accuracy of 80.5%. The highest prediction accuracies are obtained for Openess to Experience, Neuroticism, and Extraversion, reaching above 84%, while for Agreeableness and Conscientiousness we only obtained roughly around 77%. Overall, the prediction accuracy of the system is higher than that of any other state-of-the-art method tested on the same database. Another significant finding is that we determined several relationships between the prediction with high accuracy of specific FMM personality traits and the handwriting features analyzed which can be further exploited to improve the accuracy of the system. The accuracy of the system can also be further improved either by analyzing other handwriting features together with the seven ones already analyzed in our study or grouping these features based on the relevant information they offer in this task and filter out the irrelevant ones for each of the five personality traits. This will be the direction of our future research.

The proposed system computes the results in no more than 90 s which makes it faster than the current ways of determining personality traits through extensive self-report questionnaires, usually more cumbersome and time-consuming to fill in and involving more effort from both subject’s and psychologist’s side which will have to post-process the questionnaire results and evaluate the five personality traits; this shows that our current approach could be used as an attractive, faster, and easier to use alternative to these commonly used personality evaluation techniques.

Abbreviations

Average Absolute Relative Error

Artificial neural networks

Autism spectrum disorder

Controlled Dataset

Five-Factor Model–Neural Network

Five-Factor Model

Handwriting Map

Hidden Markov Model

K-nearest neighbors

Myers-Briggs Type Indicator

Minnesota Multiphasic Personality Inventory

Modified National Institute of Standards and Technology

Neural network

Parkinson’s disease

Random Dataset

Random Access Memory

Support Vector Machines

Stroke Width Transform

Vertical Projection Profile

R Plamondon, Neuromuscular studies of handwriting generation and representation. International Conference on Frontiers in Handwriting Recognition (ICFHR), 261 (2010) Kolkata, November 2010

Y Tang, X Wu, W Bu, Offline text-independent writer identification using stroke fragment and contour-based features. 2013 IEEE International Conference on Biometrics (ICB), 1–6 (June 2013)

M Naghibolhosseini, F Bahrami, A behavioral model of writing. International Conference on Electrical and Computer engineering (ICECE), 970–973 (December 2008)

HN Champa, KR Anandakumar, Automated human behavior prediction through handwriting analysis. 2010 First International Conference on Integrated Intelligent Computing (ICIIC), 160–165 (August 2010)

B Fallah, H Khotanlou, in Artificial Intelligence and Robotics (IRANOPEN) . Identify human personality parameters based on handwriting using neural networks (April 2016)

Google Scholar  

Z Chen, T Lin, Automatic personality identification using writing behaviors:an exploratory study. Behav Inform Technol 36 (8), 839–845 (2017)

Article   Google Scholar  

I Siddiqi, C Djeddi, A Raza, L Souici-Meslati, Automatic analysis of handwriting for gender classification. Pattern. Anal. Applic. 18 (4), 887–899 (November 2015)

Article   MathSciNet   Google Scholar  

S Maadeed, A Hassaine, Automatic prediction of age, gender, and nationality in offline handwriting. EURASIP Journal on Image and Video Processing 2014 , 10 (December 2014)

G Luria, S Rosenblum, A computerized multidimensional measurement of mental workload via handwriting analysis. Behav. Res. Methods 44 (2), 575–586 (June 2012)

I Zaarour, L Heutte, P Leray, J Labiche, B Eter, D Mellier, Clustering and Bayesian network approaches for discovering handwriting strategies of primary school children. Int. J. Pattern Recognit. Artif. Intell. 18 (7), 1233–1251 (2004)

R Sudirman, N Tabatabaey-Mashadi, I Ariffin, Aspects of a standardized automated system for screening children’s handwriting. First international conference on Informatics and Computational Intelligence (ICI) , 48–54 (December 2011)

J Gorbova, I Lusi, A Litvin, G Anbarjafari, Automated screening of job candidate based on multimodal video processing. Computer Vision and Pattern Recognition Workshops (CVPRW) (2017) IEEE Conference on , July 2017

G Luria, A Kahana, S Rosenblum, Detection of deception via handwriting behaviors using a computerized tool: Toward an evaluation of malingering. Cogn. Comput. 6 (4), 849–855 (December 2014)

TLP Tang, Detecting honest People’s lies in handwriting. J. Bus. Ethics 106 (4), 389–400 (April 2012)

SB Bhaskoro, SH Supangkat, An extraction of medical information based on human handwritings. 2014 International Conference on Information Technology Systems and Innovation (ICITSI), 253–258 (November 2014)

Drotar, P., Mekyska, J., Smekal, Z., Rektorova, I., Prediction potential of different handwriting tasks for diagnosis of Parkinson’s, 2013 E-Health and Bioengineering Conference, Pages 1–4, November 2013.

N Grace, PG Enticott, BP Johnson, NJ Rinehart, Do handwriting difficulties correlated with core symptomology, motor proficiency and attentional behaviors. Journal of Autism and Developmental Disorders , 1–12 (January 2017)

WL Lee, K-C Fan, Document image preprocessing based on optimal Boolean filters. Signal Process. 80 (1), 45–55 (2000)

Article   MATH   Google Scholar  

JG Leu, Edge sharpening through ramp width reduction. Image Vis. Comput. 18 (6), 501–514 (2000)

SCF Lin et al., Intensity and edge based adaptive unsharp masking filter for color image enhancement. Optik - International Journal for Light and Electron Optics 127 (1), 407–414 (2016)

R Legault, CY Suen, Optimal local weighted averaging methods in contour smoothing. IEEE Trans. Pattern Anal. Mach. Intell. 18 , 690–706 (July 1997)

Y Solihin, CG Leedham, Integral ratio: a new class of global thresholding techniques for handwriting images. IEEE Trans. Pattern Anal. Mach. Intell. 21 , 761–768 (Aug. 1999)

Kai Chen, Fei Yin, Cheng-Lin Liu, Hybrid page segmentation with efficient whitespace rectangles extraction and grouping, Document Analysis and Recognition (ICDAR) 2013 12th International Conference on , pp. 958–962, 2013.

V Papavassiliou, T Stafylakis, V Katsouro, G Carayannis, Handwritten document image segmentation into text lines and words. Pattern Recogn. 43 , 369–377 (2010)

Costa, P.T. Jr., McCrae, R.R., Revised NEO Personality Inventory (NEO-PI-R) and NEO Five-Factor Inventory (NEO-FFI) manual, Psychol. Assess. Resources, Odessa, FL, 1992.

BW Roberts, D Mroczek, Personality trait change in adulthood. Curr. Dir. Psychol. Sci. 17 (1), 31–35 (2008)

M Jokela, C Hakulinen, A Singh-Manoux, M Kivimaki, Personality change associated with chronic diseases: pooled analysis of four perspective cohort studies. Psychol. Med. 44 , 2629–2640 (2014)

SJ Karau, RR Schmeck, AA Avdic, The big five personality traits, learning styles, and academic achievement. Journal on Personality and Individual Differences 51 (4), 472–477 (September 2011)

Morris, R. N., Forensic Handwriting Identification: Fundamental Concepts and Principles, 2000.

K Amend et al., Handwriting Analysis: The Complete Basic Book (Borgo Press, San Bernardino, California, 1981)

MB Menhaj, F Razzazi, A new fuzzy character segmentation algorithm for Persian/Arabic typed texts. International Conference on Computational Intelligence, Fuzzy Days 1999 : .Computational Intelligence, 151–158 (1999)

R Coll, A Fornes, J Llados, Graphological analysis of handwritten text documents for human resources recruitment. 12th International Conference on Document Analysis and Recognition, 1081–1085 (July 2009)

EM Hicham, H Akram, S Khalid, Using features of local densities, statistics and HMM toolkit (HTK) for offline Arabic handwriting text recognition. Journal of Electrical Systems and Information Technology 4 (3), 387–396 (2017)

B Epshtein, E Ofek, Y Wexler, Detecting text in natural scenes with stroke width transform. Computer Vision and Pattern Recognition (CVPR), 2010 IEEE Conference on (August 2010)

L Deng, The MNIST database of handwritten digit images for machine learning research [best of the web]. IEEE Signal Process. Mag. 29 (6), 141–142 (November 2012)

L Xiaoyuang, Q Bin, W Lu, A new improved BP neural network algorithm. Second International Conference on Intelligent Computation Technology and Automation, 19–22 (October 2009)

Download references

Availability of data and materials

Data is not shared publicly. Please contact the author for data requests.

Author information

Authors and affiliations.

Department of Telecommunications, University “Politehnica” of Bucharest, 1-3 Iuliu Maniu Blvd, 06107, Bucharest 6, Romania

Mihai Gavrilescu & Nicolae Vizireanu

You can also search for this author in PubMed   Google Scholar

Contributions

MG contributed to the state-of-the-art research, implementation of the neural network-based testbed and methods employed in Scala using Spark library, testing the proposed architecture and discussion around the results and conclusions. NV contributed to the discussion around the results obtained and conclusions. Both authors read and approved the final manuscript.

Corresponding author

Correspondence to Mihai Gavrilescu .

Ethics declarations

Competing interests.

The authors declare that they have no competing interests.

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License ( http://creativecommons.org/licenses/by/4.0/ ), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.

Reprints and permissions

About this article

Cite this article.

Gavrilescu, M., Vizireanu, N. Predicting the Big Five personality traits from handwriting. J Image Video Proc. 2018 , 57 (2018). https://doi.org/10.1186/s13640-018-0297-3

Download citation

Received : 20 April 2018

Accepted : 25 June 2018

Published : 10 July 2018

DOI : https://doi.org/10.1186/s13640-018-0297-3

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Neural networks
• Handwriting analysis
• Personality classification
• Feature classification

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings

Preview improvements coming to the PMC website in October 2024. Learn More or Try it out now .

• Advanced Search
• Journal List
• HHS Author Manuscripts

How Universal Is the Big Five? Testing the Five-Factor Model of Personality Variation Among Forager–Farmers in the Bolivian Amazon

Michael gurven.

Department of Anthropology, University of California, Santa Barbara

Christopher von Rueden

Maxim massenkoff, hillard kaplan.

Department of Anthropology, University of New Mexico

Marino Lero Vie

Tsimane Health and Life History Project, San Borja, Beni, Bolivia

Associated Data

The five-factor model (FFM) of personality variation has been replicated across a range of human societies, suggesting the FFM is a human universal. However, most studies of the FFM have been restricted to literate, urban populations, which are uncharacteristic of the majority of human evolutionary history. We present the first test of the FFM in a largely illiterate, indigenous society. Tsimane forager–horticulturalist men and women of Bolivia ( n = 632) completed a translation of the 44-item Big Five Inventory ( Benet-Martínez & John, 1998 ), a widely used metric of the FFM. We failed to find robust support for the FFM, based on tests of (a) internal consistency of items expected to segregate into the Big Five factors, (b) response stability of the Big Five, (c) external validity of the Big Five with respect to observed behavior, (d) factor structure according to exploratory and confirmatory factor analysis, and (e) similarity with a U.S. target structure based on Procrustes rotation analysis. Replication of the FFM was not improved in a separate sample of Tsimane adults ( n = 430), who evaluated their spouses on the Big Five Inventory. Removal of reverse-scored items that may have elicited response biases produced factors suggestive of Extraversion, Agreeableness, and Conscientiousness, but fit to the FFM remained poor. Response styles may covary with exposure to education, but we found no better fit to the FFM among Tsimane who speak Spanish or have attended school. We argue that Tsimane personality variation displays 2 principal factors that may reflect socioecological characteristics common to small-scale societies. We offer evolutionary perspectives on why the structure of personality variation may not be invariant across human societies.

The five-factor model (FFM) is a widely accepted construct describing personality variation along five dimensions (i.e., the Big Five): Extraversion, Openness, Conscientiousness, Neuroticism, and Agreeableness. Many researchers have argued that the structure of the FFM is a “biologically based human universal” that transcends language and other cultural differences ( Bouchard & Loehlin, 2001 ; McCrae & Costa, 1997 ; Wiggins & Trapnell, 1997 ; Yamagata et al., 2006 ). Cross-cultural tests of the FFM in over 50 societies across six continents have supported the existence and universality of the FFM ( McCrae, 2002 ; McCrae, Terracciano, & 78 Members of the Personality Profiles of Cultures Project, 2005 ; Schmitt et al., 2007 ). A universal structure suggests uniform covariance among traits in humans despite vastly different culture, history, economy, social life, ideology, and every other form of cultural and behavioral expression. The Big Five structure is even notable in captive chimpanzees, based on ratings by zoo employees ( King & Figueredo, 1997 ).

Despite the increasing consensus supporting the FFM, a five-factor structure does not robustly emerge everywhere, and some researchers have posited more than five personality factors within certain populations (e.g., Cheung & Leung, 1998 ; Lee & Ashton, 2004 ); however, these additional factors can often be subsumed under one of the Big Five factors ( Guanzon-Lapeña, Church, Carlota, & Katigbak, 1998 ). Thus, the FFM has yet to be robustly falsified, at least in literate, industrialized societies. If the FFM is a human universal and represents a “solid beginning for understanding personality everywhere” ( McCrae & Costa, 1997 , p. 515), it should replicate everywhere and under a broad range of environments and populations. To date, the FFM has yet to be tested in an indigenous, preliterate society. The vast majority of samples from cross-cultural studies are often urban students, glibly referred to as western, educated, industrialized, rich, democratic (WEIRD) populations ( Henrich, Heine, & Norenzayan, 2010 ). Despite the wide range of cultures and languages where the FFM has been tested, WEIRD populations might show a similar personality structure if trait covariance is an artifact of living in large urban, literate populations. There are important reasons for assessing the validity of the FFM in an indigenous, preliterate society. First, human psychological adaptations likely evolved in the ancestral context of a hunting and gathering lifestyle with a social life characterized by frequent face-to-face interactions, largely with kin. Although pure hunter–gatherers are exceedingly rare, many groups maintain traditional lifestyles and share many social and economic characteristics with hunter–gatherers. Testing the FFM in these populations would be particularly valuable for assessing the universality of the FFM. In the past, empirical patterns observed in WEIRD populations and assumed to be human universals have been contradicted (or at least qualified) by observations in small-scale societies ( Henrich et al., 2010 ). To date, no test of the FFM has ever been conducted among a small-scale population of foragers, farmers, or herders.

Second, the existence of the FFM is an inductively derived success of personality psychology, but to date, no extensive theory exists that can generate the FFM from first principles. There are no a priori reasons for expecting a particular number of trait dimensions or within-trait and intertrait correlations, although post hoc explanations of empirical regularities have been made (e.g., Denissen & Penke, 2008 ; Nettle, 2010 ). Thus, when the FFM receives less consistent support, as in several non-Western countries (e.g., Piedmont, Bain, McCrae, & Costa, 2002 ; Schmitt et al., 2007 ; Triandis, 1997 ), a common response from FFM advocates is to argue that methodological issues prevent FFM replication. However, without a comprehensive theory of personality formation, it is unclear whether different socioecological environments should generate veritable differences in personality structure in the first place. Are the tenuous results in non-Western societies genuine or artifactual?

We provide the first test of personality structure among an indigenous, largely illiterate population: the Tsimane forager– horticulturalists of lowland Bolivia. We use a Spanish translation of the Big Five Inventory, a widely used metric of the FFM first developed by Benet-Martínez and John (1998) . Our null prediction is that the Big Five should replicate in the Tsimane population. If certain features, such as literacy and education, are important for generating the Big Five pattern, we might find that the Big Five does not replicate among Tsimane. However, we should expect to find the Big Five structure to replicate among more educated and literate Tsimane. We test the validity of the five-factor model by assessing (a) internal reliability of each factor, (b) external validity of the factors, (c) 1-year test–retest factor correlations, (d) whether the FFM is generated from exploratory factor analysis, (e) whether confirmatory factor analysis supports the FFM, and (f) whether Procrustes rotation to a U.S-based sample indicates similar FFM structure. We determine whether the FFM is better replicated with (g) stratification of the sample into subgroups that might differ in familiarity with testing procedures, performance, and self-reflection (age, sex, schooling, and Spanish fluency), (h) selective removal of least internally consistent items, (i) selective removal of items that evidence socially desirable responding (i.e., highly positive or negative response scores), (j) correction for acquiescence bias (i.e., a tendency of subjects to affirm personality descriptors read to them), or (k) evaluation of a separate sample of subjects asked to evaluate the personality of their spouses. Peer-reported personality may improve internal reliability of the Big Five ( McCrae et al., 2005 ).

Despite our rigorous set of tests and analyses, we do not find strong, consistent support for the Big Five. We instead find evidence of factor structure consistent with a “Big Two” oriented around prosociality and industriousness. Our findings put the universality of the FFM into question but, more important, heighten the need to develop models of how low-order traits should be coordinated to assemble into higher order factors, given cultural and socioecological variability.

The paper is organized into five sections. Section 1 provides an overview of cross-cultural studies of the FFM in order to contextualize the value of the current study. Section 2 briefly describes the Tsimane population. Section 3 discusses our methods, and Section 4 presents our results. Section 5 interprets our results and discusses personality and the FFM in small-scale indigenous societies.

Cross-Cultural Studies of the Big Five

The FFM has been assessed with both etic and emic approaches. In etic studies, a previously identified personality structure is applied in a different culture or context; in emic approaches, a personality structure is indigenously derived with a sampling of the target culture’s personality descriptors.

The FFM was derived in English using a lexical (emic) approach, which assumes that all relevant personality descriptors are found in a group’s vocabulary ( Digman, 1990 ; Goldberg, 1990 ; John, 1990 ). Although early research in personality structure yielded many competing constructs to describe personality variation, the FFM has emerged as the most widely accepted model ( Peabody & De Raad, 2002 ). The FFM has since been tested in many countries and in numerous languages with the Revised NEO Personality Inventory (NEO-PI–R) ( Costa & McCrae, 1992 ) and the Big Five Inventory (BFI) ( Benet-Martínez & John, 1998 ) protocols. Even a nonverbal protocol has confirmed the generalizability of the FFM in cross-cultural context ( Paunonen, Ashton, & Jackson, 2001 ).

Across cultures, etic studies have generally replicated the FFM (NEO-PI–R: McCrae, 2002 ; BFI: Schmitt et al., 2007 ), and factor scales show high internal reliability; however, Extraversion and Agreeableness are sometimes sensitive to “cultural effects” and are not always clearly differentiated ( Ortiz et al., 2007 ; Rolland, 2002 ). As a result, McCrae, Costa, Del Pilar, Rolland, and Parker (1998) have suggested that a universal FFM consists of the first three factors and an “interpersonal circumplex”—which subsumes elements of Extraversion and Agreeableness factors based on Procrustes analysis ( Rolland, 2002 ).

Among emic studies, an Openness factor is not consistently extracted ( De Raad, 1994 ; Di Blas & Forzi, 1998 ; Szirmák & De Raad, 1994 ). Furthermore, several emic studies have consistently yielded more than five factors ( Almagor, Tellegen, & Waller, 1995 ; Benet-Martínez & Waller, 1997 ). In China, Cheung and Leung (1998) have identified a “tradition” factor independent of the Big Five. However, results from emic studies do not always match the results from etic studies of the same population. For example, in Italy, studies using translated inventories have identified a Neuroticism factor ( Caprara, Barbaranelli, Borgogni, & Perugini, 1993 ; Perugini & Leone, 1996 ), but emic studies have not ( Caprara & Perugini, 1994 ; Di Blas & Forzi, 1998 ). Openness and Neuroticism are more robustly established in etic studies than in emic studies, which has led to a growing consensus that lexical approaches underlying emic studies are not comprehensive ( Church & Lonner, 1998 ; Rolland, 2002 ). As McCrae and Costa (1997) concluded, “It is simply not the case that all personality traits are encoded as adjectives … lexical studies confound differences in personality structure with differences in personality language” (p. 510).

In cross-cultural studies, reliability of the FFM has been highest in developed countries. In Allik and McCrae (2004) and Schmitt et al. (2007) , sample populations were predominantly college students and were often bilingual. In developing countries, the FFM has met with less success; whether this is due to methodological problems or to actual differences in personality structure remains to be determined. Methodological differences may arise due to translations not being equivalent, lack of item relevance in the local culture, differences in subject response styles, unfamiliarity with the test format, and unrepresentative samples ( Paunonen & Ashton, 1998 ).

In Schmitt et al. (2007) , internal consistency of factor items based on Cronbach’s alpha was sufficiently high in South American samples, with each country averaging above the standard benchmark of 0.70. However, several African countries fared worse: Average Cronbach’s alphas for Morocco, Tanzania, Ethiopia, and Congo were 0.62, 0.59, 0.48, and 0.48, respectively. Despite low internal consistency, the African and South American samples showed high levels of congruence with the American normative factor structure under Procrustes rotation ( Schmitt et al., 2007 ). However, of the seven countries in Africa reported in Schmitt et al. (2007) , six were administered the BFI in English, and four had samples restricted to college students. Similarly, the five South American countries in the study (including Bolivia) contained only college students.

Reliability is sometimes improved in studies that rely on third-party observer reports rather than self-reports. In a large cross-cultural study of this type in 50 different societies, McCrae et al. (2005) asked college students to give observer ratings on the NEO-PI–R for persons of all ages they knew well. Roughly 5% of the Cronbach alphas were lower than 0.70, with this 5% concentrated primarily in the samples from developing countries. Although relying on observer ratings helped improve internal consistency, it did not eliminate potential problems of evaluative bias common to self-report data in developing societies. For example, Openness did not cleanly emerge in Nigeria. McCrae et al. (2005) concluded that “it is possible that there is a minority of cultures in which the [FFM] structure is not found” (p. 552).

To our knowledge, only two studies have focused explicitly on ethnic populations in the developing world. Piedmont et al. (2002) tested the NEO-PI–R among the Shona, a sub-Saharan society in Zimbabwe. Within this mixed rural and suburban sample (predominantly college students bilingual in English and their native Shona), the average internal consistency for the five factors was 0.77, higher than for the African samples in Schmitt et al. (2007) . However, Openness produced a low reliability of 0.64, and only five of the 30 NEO-PI–R facets produced reliabilities above 0.60. Factor congruence with the American normative structure was high at 0.89, but only 15 facets produced congruence coefficients higher than 0.90. These results were obtained with the Shona language version of the NEO-PI–R; the English version of the test showed slightly higher reliability and congruence. Schmitt et al. identified translation problems as the main factor contributing to the less than ideal fit to the FFM: The Shona language lacks words equivalent to some of the English terms in the NEO-PI–R.

Alvergne, Jokela, and Lummaa (2010) administered the English Mini-Markers Big Five Inventory ( Thompson, 2008 ) in four agricultural Senegalese communities, among individuals with diverse ages and with low levels of education. The subsistence focus on cash cropping and the low fertility rate (5 births per woman) are not characteristic of more traditional human societies lacking agriculture and practicing natural fertility. The sample size was quite small ( n = 65 families), and the Mini-Markers Inventory used has not been validated among non-English speakers. After removal of hard-to-translate items and further shortening of the survey for brevity, the administered version of the BFI included only 27 items. Alvergne et al. retained about half of those items for analysis, with most factors based on only two or three adjectives. Reliability among these factors was still low, averaging 0.64.

Study Population

The Tsimane are forager–horticulturalists of central lowland Bolivia, located along the Maniqui, Quiquibey, Apere, and Matos Rivers and in adjacent forests of the Beni Department. Although families may spend weeks or months on hunting or fishing trips or cultivate fields some distance from their primary house in settled villages, the Tsimane are semisedentary and live in communities ranging from 30 to 500 individuals. Their population is estimated at 10,000 and is dispersed among over 90 villages. They cultivate plantains, rice, corn, and sweet manioc in small swiddens and regularly fish and hunt for meat. These foods together provide over 90% of the calories in the diet, with the remainder coming mainly from trade with itinerant merchants. Polygyny occurs at low frequencies (~5%) and is concentrated in more remote communities ( Gurven, Winking, Kaplan, von Rueden, & McAllister, 2009 ). Exclusive priority of access for individuals or small groups to certain rights and resources is minimal, but land close to village centers is de facto privately owned. More extensive ethnographic background can be found in Chicchón (1992) , Huánca (1999) , and Schniter (2009) .

Since the mid-20th century, the Tsimane have come into greater contact with modernizing influences. In Tsimane villages, especially those located near the town of San Borja (population ~25,000), incipient cattle ownership, wage labor with loggers and farmers, and produce sales to local markets are on the rise. Many Tsimane now have minimal access to health care through the services of a health post, a hospital in San Borja, and the Tsimane Health and Life History Project, but mortality rates remain high, particularly among infants. Approximately 20% of offspring never reach age 5 ( Gurven, Kaplan, & Zelada Supa, 2007 ). The Tsimane rarely use modern contraceptives; the total fertility rate is very high (~9 births per woman), and so the population growth rate is high (3.6% per year). Many Tsimane villages now have access to public schooling for their children taught largely by bilingual Tsimane teachers trained by local missionaries. Several secondary schools now exist in larger villages, and young Tsimane adults are starting to become high school graduates. However, the overall adult literacy rate remains low, at 25%. Fluency in the native Tsimane language is universal, and only 40% of adults are moderately fluent in Spanish. The Tsimane language is an isolate, together with Mosetene, and it is unrelated to the dominant indigenous languages of Bolivia.

Tsimane live in extended family clusters, within which occur the majority of food and labor sharing. Although social and cooperative in daily interactions with village co-residents, Tsimane families value their autonomy. Groups of family clusters compose villages, which were given formal geographic boundaries only in the late 20th century and lack a strong sense of identity ( Gurven, Zanolini, & Schniter, 2008 ). Village residents elect chiefs to organize community meetings and to represent their interests to outside political bodies, but chiefs lack any substantial authority, tend to have short tenure, and often are unable to effectively organize people for collective action ( Gurven & Winking, 2008 ; von Rueden, Gurven, & Kaplan, 2008 ). In the event of interpersonal conflict, Tsimane often “vote with their feet” by moving to other villages.

Tsimane often describe each other in valent terms, with judgments of good ( jäm’si ) and bad ( jam jäm’si or a’chis ) applying to numerous domains. Maintaining friendly relations ( jäm’yity muntyi ), being easygoing ( chuchuijtyi ), and avoiding direct confrontation and expression of anger ( chij facoij ) are viewed as proper ways of behaving and are ingrained in Tsimane culture. In their descriptions of others, Tsimane recognize the persistence of particular traits in individuals over time. Someone who speaks freely ( chij peyaquity ) but not too much or in a gossiping way ( chij peyacsity ) is a valued social partner, and jokesters are also recognized and viewed positively ( chij shevinyity ). Happy, cheerful individuals ( majoijbäyis ) are contrasted with serious, quiet individuals ( futy’dyety ) or those who are easily annoyed ( achiyity ). Other negative traits commonly described refer to those who react rapidly, usually in a bad way ( che’chei’si ), those who brag ( va’bunyis ), and those who are lazy ( shoyijyi’tyi or jamyedyedyetyi ). Laziness is often contrasted with demonstration of strong work effort ( setyi or chij carijtaqui ) and generosity in helping others ( chij notacsity ).

We administered a personality questionnaire based on the Big Five Inventory (BFI), a widely used 44-item metric of the five-factor model. The Spanish version of the BFI, previously validated by Benet-Martínez and John (1998) , was translated into the Tsimane language by two bilingual Tsimane research assistants (Marino Lero Vie [MLV] and Feliciano Cayuba Claros) and Michael Gurven (MG). As a test of the accuracy of the translation, the Tsimane questionnaire was then back-translated into Spanish by a different translator, and discussions among the three bilingual Tsimane and MG ensued until a workable translation was found that captured the essence of each item. Due to limitations of Tsimane vocabulary, several items required a definitional phrase in the local idiom rather than relying on a single word to capture the right meaning. In these cases, either an exact word did not exist or, taken out of context, the word could be misconstrued. For example, Item 31 (“is clever and analytical”) was translated as Mi buty chij cave'jedye judyeya jäm' yu' ban mi (literally, “Knows how to ‘see’ things and can make things turn good”), because the Tsimane word for “smart” reflects the state of being knowledgeable. Item 32 (“radiates enthusiasm”) was translated as Mi buty fer ma'je' ji'cave' jun'si chuc mi ma'je (literally, “You really show to others whatever it is you want” [to show]) because there are no Tsimane words for “radiate” or “enthusiasm.” Due to the lack of any word for “art” in Tsimane, Item 44 (“few artistic interests”) was translated more descriptively as “someone who does not like to play music, sing, tell stories, or draw.” Those are the main forms of artistic expression in Tsimane society. When necessary, translating the whole concept rather than the literal words enabled us to circumvent translation problems reported by other cross-cultural studies of the FFM (e.g., Piedmont et al., 2002 ). Only one item from the original BFI was removed (Item 30: “has an active imagination”) due to the inability to find a suitable expression to explain the concept in a manner that was consistently understood by Tsimane subjects. This item, alone among the BFI items, was found to be understood differently by bilinguals when presented in Spanish versus English, suggesting it should be revised or omitted from the BFI in the context of cross-cultural studies ( Ramírez-Esparza, Gosling, Benet-Martínez, Potter, & Pennebaker, 2006 ). Thus, the final Tsimane BFI instrument includes 43 items.

The Tsimane BFI was administered to 632 adults from 28 villages during the period January 2009 to December 2010. The sample was 48% female, the average age was 47 years (range = 20–88 years, SD = 14.4), and the average years of formal education was 1.2 years (range = 0–12 years, SD = 2.2). The age, years of formal education, and Spanish proficiency of all subjects were ascertained from demographic interviews (see Gurven et al., 2007 ). The Tsimane BFI was conducted verbally in a private location by a bilingual Tsimane research assistant (MLV) trained in the administration of anthropological and psychological interviews. As in the English version of the BFI, responses were given on a translated scale where 1 corresponds to strongly disagree and 5 corresponds to strongly agree . Subjects were first given a quick tutorial and comprehension test on the use of the scale, after which all subjects showed clear evidence of understanding the scale and the task at hand. The scale, depicted on a piece of cardboard placed in front of the subject, included drawings to help facilitate understanding. Five drawings of a person accompanied the five numbers on the scale; the drawings revealed more and more of the person as the scale ascended: a drawing of just a person’s legs accompanied 1 and a drawing of the whole body accompanied 5. Although many respondents were previously unfamiliar with Likert-type scales, few were new to formal interviews because of their extensive participation in the Tsimane Health and Life History project we have maintained continuously since 2002 (see http://www.unm.edu/~tsimane/ ). Indeed, our decade-long presence in the area has helped to establish trusting, collaborative relationships among study subjects.

After the interview, MLV used the same 5-point scale to rate respondents on four variables based on his observations during the fifteen or so minutes of the BFI interview together with an additional 30 minutes spent conducting a separate interview (on economic production and sharing): the extent to which the subject was talkative, shy, smiling and/or joking, and easily distracted. These were added to help gauge external validity of the FFM instrument. MLV performed multiple test runs in order to ensure consistency in his observations.

None of our interviews produced missing items. Thirty-four subjects (53% female) were interviewed twice, each interview roughly a year apart (average 14.2 ± 2.6 months), providing a test of response stability. The average age of this subsample is 52 years.

In addition to conducting our first-person interviews, we asked 430 Tsimane adults to rate their spouses on the Tsimane BFI. These interviews were conducted during the period from March 2011 to February 2012. The sample of spouses who were rated was 50% female, and the average age was 52 years (range = 16–89 years, SD = 11.6). The protocol did not differ from the self-report protocol except that with each item of the BFI verbalized to the raters, subjects were reminded to evaluate their spouse. The self-report and spouse-report samples overlap for 66 individuals (46% female; average age = 52 years). Although the spouse-report sample by definition excludes unmarried individuals, we do not expect significant differences across the samples due to marital status: Only 26 of the 632 adults in the self-report sample were single at the time of data collection.

Internal Reliability

We first test the reliability of each of the Big Five factors. The Cronbach’s alpha measures of internal reliability, factor means, ranges, and standard deviations are given in Table 1 . All items phrased in reverse (e.g., the Extraversion item “is shy”) were reverse scored prior to calculation of these statistics. Although the distributions of subjects’ scores on the Big Five factors do not conform to a normal distribution according to the Shapiro–Wilk test, the distributions do not exceed skew or kurtosis values of ± 1. Extraversion, Agreeableness, Conscientiousness, and Openness show moderate internal reliability (Cronbach’s α = 0.63, 0.58, 0.69, and 0.54, respectively), and Neuroticism shows low reliability (0.31).

Mean Response Score, Score Ranges and Standard Deviations, and Internal Reliability (Cronbach’s Alpha) for the Five Factors

Internal Reliability by Age, Sex, Education, and Spanish Fluency

We next examine whether internal reliability differs by age, sex, formal education, and Spanish fluency. If schooled adults are more familiar with testing and if Spanish speakers are more familiar with other ideas and cultures in a way that may promote self-reflection, then their item responses within factors might be more consistent than responses from unschooled or monolingual Tsimane speakers. Subjects were divided into several subgroups: those older and younger than 44 years (the median age), men and women, those with and without any formal schooling, and those who do or do not speak Spanish. Although internal reliability of several of the Big Five improves within particular subgroups, no subgroup shows consistent improvement across all of the Big Five (see Table 2 ). Averaged across the Big Five, differences in reliability between complementary subgroups (e.g., old vs. young) were close to zero. Extraversion and particularly Openness show higher internal reliability among men, the young, the educated, and those who speak Spanish. Agreeableness and Conscientiousness produce the opposite result.

Internal Reliability Based on Cronbach’s Alpha for Subgroups of Self-Report Sample

Removing Potentially Problematic Items and Correcting for Acquiescence Bias

We consider the possibility that despite our efforts at repeated translation and back-translation, certain items may still have been interpreted differently by subjects from their intended meaning. If certain items are driving the low reliability scores, we might expect them to load weakly on each factor. In an attempt to address this potential problem, we first drop the least reliable item (i.e., the item whose removal would most increase factor internal reliability) from each of the Big Five and recalculate Cronbach’s alpha. Extraversion and Conscientiousness now surpass the standard benchmark of 0.70, and internal reliability for Agreeableness and Openness improve but remain suboptimal. The reliability for Neuroticism remains quite low even after removal of the least reliable item (see Table 1 ). The least internally reliable items include, for Agreeableness, Item 22 (“is sometimes ill-mannered with others”); for Conscientiousness, Item 42 (“gets distracted easily”); for Extraversion, Item 6 (“is reserved”); for Neuroticism, Item 35 (“remains calm in difficult situations”); and for Openness, Item 12 (“likes routine”). Further removal of the weakest remaining item from each factor did not bring Agreeableness, Neuroticism, or Openness to acceptable levels of reliability.

The first and second least reliable items within each of the Big Five are all items that are reverse scored. This suggests these items may have been differentially susceptible to socially desirable responding. Alternatively, a low covariation among true- and reverse-scored items within each of the Big Five could arise through acquiescence bias, which is any tendency of individuals to respond affirmatively to questions posed them. We remove all reverse-scored items and recalculate Cronbach’s alpha for each of the Big Five. This eliminates 16 of the 43 items. Agreeableness, in addition to Extraversion and Conscientiousness, now produces acceptable internal reliability. The reliabilities for Neuroticism and Openness remain low (see Table 1 ).

We next assess internal reliability by removing other items that may have prompted socially desirable responding. These are items with high or low mean response values. Given the self-report nature of the BFI instrument, especially to a third-party (albeit neutral) Tsimane assistant, it may be that an individual less familiar with interviews (a) is uncomfortable conveying self-ratings for traits deemed highly negative or (b) gives biased responses for highly positive traits when speaking to another Tsimane (or even to him- or herself). We therefore remove items with mean response scores less than two or greater than four. This eliminates nine of the 43 items: two with strong disagreement (Item 2: “tends to be critical”; Item 13: “starts disputes with others”) and seven with strong agreement (Item 3: “is meticulous about work”; Item 10: “has diverse interests”; Item 11: “energetic”; Item 23: “is inventive”; Item 26: “worries about things”; Item 35: “maintains calm in difficult situations”; Item 37: “is considerate and friendly with everyone”). This exercise modestly increases internal reliability for Neuroticism yet decreases reliability for Agreeableness, Openness, Extraversion, and Conscientiousness (see Table 1 ). Thus, with this manipulation, none of the Big Five surpass a Cronbach’s alpha score of 0.70. It is noteworthy to mention that for at least five of these eliminated items, means distant from 3 are unsurprising and mesh with our expectations based on 12 years of experience living with Tsimane.

Finally, we attempt to correct for acquiescence bias not by removing problematic items but according to the method described in Hofstee, Ten Berge, and Hendriks (1998) . First, we average the response scores for each subject for 15 BFI item pairs with opposite implications for personality ( Soto, John, Gosling, & Potter, 2008 ). Second, we generate an acquiescence index by calculating the difference between each average and the scale midpoint. Third, we subtract each subject’s acquiescence score, whether positive or negative, from his or her responses. The average acquiescence score across the 632 subjects is 0.23 ( SD = 0.29), which is 5.84% of the scale range. Acquiescence in Western subjects is of a similar magnitude: Rammstedt, Goldberg, and Borg (2010) reported an average acquiescence score on the BFI of 0.11 ( SD = 0.28) for German adults with a high degree of formal education and an average score of 0.25 ( SD = 0.38) for those with little or no formal education. Among the Tsimane, correction for acquiescence bias generates acceptable internal reliability only for Conscientiousness. Internal reliability decreases significantly for Openness (see Table 1 ).

External Validity

The Big Five are correlated in expected directions with observed characteristics of subjects during interviews (see Table 3 ). Extraversion, Agreeableness, Conscientiousness, and Openness are positively correlated with smiling and negatively correlated with shyness. They also positively correlate with talkativeness and negatively correlate with distractedness, but the effect sizes are smaller. Neuroticism is positively correlated with the respondent’s shyness and negatively correlated with smiling.

Spearman Correlations of the Five Factors With Subjects’ Observed Characteristics (Self-Report Sample)

Response Stability

Test and retest responses were collected about a year apart from 34 subjects. The Tsimane average retest correlation (Spearman’s rho) is 0.431 and ranges from 0.274 ( p = .116, two-tailed) for Agreeableness, 0.370 ( p = .031) for Neuroticism, 0.420 ( p = .013) for Openness, 0.466 ( p = .005) for Conscientiousness, to 0.627 ( p < .001) for Extraversion.

Correlations Between Factors

Spearman correlations among the Big Five are presented in Table 4 . All correlations are significant at the 1% level. Neurotic individuals are less likely to be extraverted, agreeable, open, and conscientious. All other associations among other factors are positive. Extraversion is especially highly correlated with each of the other Big Five.

Spearman Correlations Between Factors (Self-Report Sample)

Note . All correlations are significant at p < .01 level.

Exploratory Factor Analysis

We perform an exploratory factor analysis (EFA) using varimax rotation and principal-components extraction to test whether our 43 BFI items inductively organize into the familiar Big Five. The unrestricted EFA results in 11 components with eigenvalues greater than one, and the eigenvalues decrease sharply after the first component (see Figure 1 ).

Scree plots for unrestricted exploratory factor analysis (self-and spouse-report samples).

Before factor rotation, the first factor explains 20.8% of the variance in the data, and the second factor explains only 5.2% of the variance. After factor rotation, this disparity is attenuated: The first factor explains 13.2% of the variance, the second explains 9.8%, and the third through fifth factors explain approximately 4.0% of the variance each. The rotated component matrix shows considerable cross-loading of items from the BFI, with no clear replication of any Big Five factor (see Table S1 of the supplemental materials ). Only the first and second factors are well defined based on the intercorrelations of items that load the highest on each factor. Cronbach’s alpha is 0.88 for the first factor, 0.83 for the second factor, and < 0.55 for subsequent factors in the unrestricted EFA. Restricting the EFA output to five factors does not noticeably improve replication of the Big Five (see Table 5 ).

Rotated Component Matrix Restricted to Five Factors (Self-Report Sample)

Note . Bolded numbers indicate on which factor items load the highest.

Stipulating a five-factor structure, we perform several EFAs with different subsets of the BFI items, with different subject subgroups, and with the data corrected for acquiescence bias. We (a) remove the 16 reverse-scored items; (b) remove items that may have prompted socially desirable or norm-conforming responses, as determined by item mean response scores of more than four or less than two; (c) transform the data to account for subjects’ degree of acquiescence bias; and (d) split the data by sex, age, schooling, and Spanish fluency. None of these manipulations clearly indicate a Big Five factor structure as determined by the rotated component matrices (see Tables S2–S12 of the supplemental materials ), and all exhibit a large first component that, prior to factor rotation, explains on average 3.2 times more of the variance in the data than the second component. Most Extraversion items load highly on the first derived factor, in addition to items from each of the other Big Five. Comparison of the items composing the derived factors ( Tables 5 , S2–S12 ) reveals a similar personality structure across most EFA subsets. Removing reverse-scored items ( Table S2 ) and correcting for acquiescence ( Table S4 ) produce factors suggestive of Agreeableness and Conscientiousness. However, many of the Agreeableness and Conscientiousness items continue to load highly on more than one factor. An EFA restricted to true-scored items from Extraversion, Agreeableness, and Conscientiousness comes closer to replicating those factors (see Table S13 of the supplemental materials ).

Confirmatory Factor Analysis

We use maximum likelihood estimation to test the fit of the self-report sample ( n = 632) to the FFM in a confirmatory factor analysis (CFA). The estimated model contains 96 free parameters, including 10 covariances among the Big Five latent variables, 38 paths from the latent variables to the observed BFI items, and 48 variances. Model fit is poor: χ 2 (850, N = 632) = 2,695.247, p < .001; root-mean-square error of approximation (RMSEA) = 0.059, 90% CI [0.056, 0.061]; comparative fit index (CFI) = 0.716; Akaike information criterion (AIC) = 2,887.247. We also perform a CFA with the 16 reverse-scored items removed, given their negative effects on internal reliability of the Big Five, particularly Extraversion, Agreeableness, and Conscientiousness. Model fit is improved but still a poor match to the data: χ 2 (314, N = 632) = 1,086.643, p < .001; RMSEA = 0.062, 90% CI [0.058, 0.067); CFI = 0.823; AIC = 1,214.643.

Procrustes Rotation

Standard protocol for assessing the comparability of personality structure across two populations involves a Procrustes rotation of sample data and estimation of factor congruence with another population that strongly displays the Big Five ( McCrae, Zonderman, Costa, Bond, & Paunonen, 1996 ; Piedmont et al., 2002 ; Schmitt et al., 2007 ). Despite our inability to reveal the Big Five using EFA or CFA, we consider the possibility that Tsimane personality structure may nonetheless be statistically similar to that in samples that typically do. We use Procrustes analysis to determine the factor congruence between our sample and a target structure, in this case a U.S. sample ( n = 2,793 college students, 64% female) from Schmitt et al. (2007) . McCrae et al. (1996) showed that Procrustes analysis is a more accurate test of replication than confirmatory factor analysis. It has since been used to successfully replicate the Big Five model within several novel samples (e.g., Piedmont et al., 2002 ; Schmitt et al., 2007 ). Congruence scores above 0.90 are indicative of good fit ( McCrae et al., 1996 ). As shown in Table 6 , Conscientiousness has the most congruence with the U.S. sample, and Neuroticism produces the least congruence. Although congruence does not improve to acceptable levels when using any of the subsamples described in previous sections, removing reverse-scored items from each of the Big Five does improve congruence (see Table 6 ). Splitting the data by age or sex does not notably improve congruence within any of the subgroups. Performing the same analysis on the loadings derived from the educated and Spanish-speaking subgroups actually decreases congruence for most factors. Removal of items with high and low average response scores and correction for acquiescence bias produce significant increases in congruence only for Neuroticism.

Procrustes Congruence With U.S. Target Structure

Comparison With Spouse Reports

Finally, we assess whether spouse-reported personality improves replication of the Big Five among the Tsimane. Internal reliability of the Big Five is lower than in the self-report sample (see Table 1 ). Cronbach’s alpha scores do not climb above 0.70 even after removal of the least reliable item within each factor, removal of reverse-scored items, removal of items with average scores more than four or less than two, and correction for acquiescence bias. The exception is Conscientiousness, which reaches acceptable internal reliability with removal of reverse-scored items.

Exploratory factor analysis using varimax rotation and principal-components extraction produces 11 factors with eigenvalues greater than one. There is less disparity in variance explained between the first and second factors than in the self-report sample (see Figure 1 ). Before factor rotation, the first factor explains 17.5% of the variance in the data and the second factor explains 10.5% of the variance. After factor rotation, the first factor explains 10.4% of the variance, the second 10.2%, the third 7.0%, the fourth 4.4%, and the fifth factor 4.2% of the variance. As with the self-report sample, the rotated component matrix shows considerable cross-loading of items from the BFI, and internal consistency is high for only the first two factors (see Table S14 of the supplemental materials ). Cronbach’s alpha is 0.85 for the first factor, 0.81 for the second factor, and < 0.65 for subsequent factors. Restricting the EFA output to five factors does not improve replication of the Big Five (see Table 7 ).

Rotated Component Matrix Restricted to Five Factors (Spouse-Report Sample)

Procrustes analysis does not indicate factor congruence with a U.S. sample that strongly displays the Big Five (see Table 6 ). Conscientiousness has the highest congruence coefficient at 0.72, and Neuroticism produces the lowest congruence coefficient at 0.38. Average congruence is lower than for the self-report sample.

We use maximum likelihood estimation to test the fit of the spouse-report data to the FFM in a CFA. The estimated model contains 96 free parameters, including 10 covariances among the Big Five latent variables, 38 paths from the latent variables to the observed BFI items, and 48 variances. Model fit is poor: χ 2 (850, N = 431) = 3,126.172, p < .001; RMSEA = 0.079, 90% CI [0.076, 0.082]; CFI = 0.523. Akaike information criteria indicate that the self-report data (AIC = 2,887.247) is a better fit than the spouse-report data (AIC = 3,404.172) to the FFM.

As we report above, only the first two factors from the self- and spouse-report samples exhibit high internal reliability in an unrestricted EFA, based on the items that load the highest on each derived factor (see Tables S1 and S14 of supplemental materials ). Given the low intercorrelations of the items within factors beyond the first two, we consider these factors poorly defined (see Tabachnick & Fidell, 2001 ). A scree test corroborates the emergence of only two well-defined factors in the spouse-report sample but is more indicative of a single factor in the self-report sample (see Figure 1 ).

Using Procrustes analysis, we test congruence between the unrestricted EFA solutions for the self- and spouse-report samples. Congruence between the second self-report factor and the first spouse-report factor is high (0.91); seven of the eight items that load the highest on the latter also load the highest on the former (see Tables S1 and S14 of supplemental materials). Congruence is also high (0.89) between the first self-report factor and the second spouse-report factor, though this is nonobvious from comparison of Tables S1 and S14 . Only four of the items that load the highest on the second spouse-report factor load the highest on the first self-report factor. However, congruential rotation takes advantage of the fact that the additional items loading highly on the first self-report factor show considerable cross-loading across the spouse-report derived factors. Subsequent factors from the self-report data produce lower congruence with the spouse-report factors, with coefficients ranging from 0.70 to 0.34.

We find significant response stability for the first two derived factors, based on the 34 individuals who self-reported their personality in 2009 and again a year later. To generate individuals’ scores on a particular derived factor, we used least squares regression. The retest correlation (Spearman’s rho) is 0.741 ( p < .001) for the first derived factor and 0.361 ( p < .036) for the second derived factor.

The items composing the first two derived factors include traits from all Big Five factors, although Extraversion and Agreeable-ness items load more highly on one factor, whereas Conscientiousness items load more highly on the other (see Table 5 and S1 of the supplemental materials ). The Spearman correlation between the two factors is 0.019 ( p = .640).

Evidence for the five-factor structure of personality among the Tsimane of Bolivia is weak. Internal reliability is generally below levels found in developed countries. The five-factor model did not cleanly emerge in any of the exploratory or confirmatory factor analyses, and Procrustean rotations did not produce strong congruence with a U.S. sample. Procrustes analysis, which is arguably the most forgiving test for replication of the FFM ( McCrae et al., 1996 ), yielded an average congruence coefficient of 0.62. This is well below the benchmark of 0.90 and considerably less than most congruence scores found in other cross-cultural applications of the Big Five ( McCrae et al., 2005 ; Schmitt et al., 2007 ).

We were able to discount several possible explanations for our results. First, we found no significant differences in structure replication after stratifying the sample by education level, Spanish fluency, sex, or age cohort. Despite research showing that education increases abstract reflection as measured by IQ (e.g., Ceci, 1991 ), educated and Spanish-speaking subsamples did not produce better replication of the Big Five among the Tsimane. Younger individuals (who are also more educated and more fluent in Spanish) were no more likely than older adults to display the Big Five. Similarly, men (who are also more educated and more fluent in Spanish) were no more likely than women to display the Big Five. These results are not surprising, in light of the fairly limited variation in Tsimane lifestyles and participation in traditional village life. Even the youngest and most educated Tsimane remain deeply embedded in traditional practices of food production and social exchange within their villages, which may partly explain why we find minimal differences in factor structure across these subsamples.

Second, removal of items with high or low average response scores did not improve replication of the Big Five relative to the full set of BFI items. Approximately one quarter of the items in the Tsimane BFI produced average responses below two or above four; these items may have elicited more socially desirable responding than other items. Studies that claim evidence for one or two higher order personality factors (e.g., Digman, 1997 ; Musek, 2007 ) have been interpreted as artifacts of socially desirable responding ( Bäckström, Björklund, & Larsson, 2008 ; McCrae et al., 2008 ). However, removal of items with low and high average response scores did not produce any closer fit to the FFM.

Third, a correction for acquiescence bias did not provide better support for the FFM. Acquiescence bias is indicated by inconsistent responding to items describing similar personality traits ( Hofstee et al., 1998 ) and has been linked with lower educational attainment ( Narayan & Krosnick, 1996 ; Rammstedt et al., 2010 ). However, our correction for acquiescence bias did not improve internal reliability of the Big Five or produce a significantly better overall fit to the FFM in EFA or Procrustes analysis.

Fourth, removal of reverse-scored items improved fit to the FFM in confirmatory factor analysis, but the fit remained poor. The reverse-scored items were the least consistent items within the Big Five, suggesting they were differentially susceptible to response biases. With the reverse-scored items removed, Extraversion, Agreeableness, and Conscientiousness just exceeded the threshold for acceptable internal reliability, and they showed clearer differentiation in exploratory factor analysis. However, items composing these factors continued to load highly on more than one factor, and Extraversion and Agreeableness items retained substantial covariation. Congruence with a U.S. target structure was higher than with our other subsamples but remained well below the benchmark of 0.90.

Fifth, we find that subjects’ personality as reported by their spouses does not support the FFM. Compared to self-report, peer report may be less influenced by response styles and has been shown to increase internal reliability among the Big Five ( McCrae et al., 2005 ; Riemann, Angleitner, & Strelau, 1997 ). Among the Tsimane, however, spouse-reported personality produced a worse fit than did self-reported data to the FFM, based on tests of internal reliability, EFA, CFA, and Procrustes congruence analysis with comparison to a U.S. target structure.

Additional evidence supports the lack of the FFM among the Tsimane. Retest correlations amongst the 34 Tsimane respondents sampled twice are significant for all Big Five factors but Agree-ableness. However, the average retest value of 0.415 is substantially lower than the ~0.65 median retest correlation for the Big Five in Western adult samples ( Costa & McCrae, 1994 ). Furthermore, Agreeableness produced the lowest retest correlation even though Neuroticism and Openness fared worse in tests of internal reliability.

We find relatively high significant correlations across the Big Five (see Table 4 ), of higher magnitude than typically found in populations where the Big Five is evident. Thus, even though we find evidence that responses to the Tsimane BFI show external validity with observed characteristics of subjects, these observations are correlated across all Big Five factors. For example, Tsimane individuals who score higher in Neuroticism are observed to be more shy and to smile less. Individuals who score higher in Extraversion are observed to be less shy and to smile more often. However, these observations of extraverts also characterize individuals who score higher in Agreeableness, Conscientiousness, and Openness. Our evidence of external validity is therefore less indicative of the FFM than other factor structures.

A valid test of the Big Five requires both that the survey items were translated accurately and that the items bear similar cultural meaning in the target society. The care with which we translated and retranslated the BFI may not preclude culture-specific interpretations of some of the items. For example, the Extraversion item “is reserved” may have been interpreted less as taciturn and more as modesty. The BFI’s reliance on dispositional terms without reference to specific situations contributes to such differences in interpretation. Successful survey instruments developed in research among Tsimane and similar groups often require concrete questions with sufficient background details (e.g., On a scale of 1–7, “how often do you hunt?” will generate more confusion and misleading responses than “In the past seven days, how many of those days did you go hunting?”). Although adding specificity to each BFI item may limit the ability to capture broader aspects of personality dimensions, it may ensure greater reliability and more meaningful responses (see Denissen & Penke, 2008 ). On the other hand, the Tsimane often speak of their peers’ personalities in the abstract (see our description of the study population), so we do not anticipate that context-specific personality items will necessarily reveal a different personality structure than manifested with our current data.

Exploratory factor analysis yields a personality structure that is largely distinct from the Big Five. Unrestricted, the factor analysis yields 11 derived factors with significant eigenvalues. When restricted to five factors, the derived factors each subsume items from at least four of the Big Five. The first derived factor is largely a mix of Extraversion and Agreeableness items and reflects a general prosocial disposition. “Reserved” and “talkative” both load positively on the first factor, but this is not necessarily contradictory. Respondents likely interpreted “reserved” as not boasting, rather than being taciturn. An egalitarian ethic among the Tsimane often curtails verbal expression of personal achievement, as is the case in many small-scale societies ( Boehm, 1999 ). The Tsimane esteem individuals who talk confidently but modestly in public settings. The Openness items “original” and “ingenious” also load positively on the first derived factor, which suggests prosocial individuals are also the most creative.

Several items from Conscientiousness sort on the second derived factor, including “efficiency,” “perseverance,” and “thoroughness.” “Energetic” and “inventive” also load highly on this factor. These items may reflect industriousness in the context of subsistence labor. Because food production labor is pooled within Tsimane extended families, it is helpful to our interpretation that “unselfishness” and “reliability as a worker” also load highly on the second factor. The third derived factor subsumes undesirable traits, whether in the context of social gatherings or labor. The fourth and fifth derived factors are more difficult to interpret and also show the least internal consistency. “Calm in tense situations” and “quiet” load positively and “quarrelsome” loads negatively on the fifth factor, which may reflect deference or reservedness in social situations. The fourth derived factor includes the items “finds fault,” “moody,” “easily distracted,” and “curious,” which is suggestive of the Western notion of (teenage) angst or, as communicated by a reviewer, an imaginative personality thwarted by a conservative society. However, these four items come from four different factors (Agreeableness, Neuroticism, Conscientiousness, and Openness, respectively).

The internal reliability of the first two derived factors in Table 5 (five-factor solution) and Table S1 (unrestricted factor solution) is high, supporting the possibility of a “Tsimane Big Two” organized according to prosociality and industriousness, as described above. These two factors show significant response stability; response stability for the first derived factor is stronger than for any of the Big Five. The spouse-report sample also produces two factors that explain more of the variance and are more internally consistent than the other derived factors. Furthermore, congruence between the self- and spouse-report samples on these first two derived factors is high. The Tsimane Big Two are therefore consistent across both self- and spouse-report samples. However, these Big Two are not the two higher order factors of Digman (1997) , characterized as stability and plasticity by DeYoung (2006) , which neatly subsume the Big Five by merging Extraversion with Openness and Agreeableness with Conscientiousness and Neuroticism. Our factors instead cut across the Big Five domains. These results are consistent with the findings of Ashton, Lee, Goldberg, and de Vries (2009) , where higher order factors emerge because lower order facets load onto multiple factors. Not only do we find that items load onto multiple factors, but the loading coefficients in our exploratory factor analyses are generally lower than those found in previous studies of the Big Five.

Our findings provide evidence that the Big Five model does not apply to the Tsimane. Our findings also bring into sharper focus past reports from developing societies where the FFM was not clearly replicated. Of the 50 countries reported in McCrae et al. (2005) , only India, Morocco, Botswana, and Nigeria produced average congruence scores less than 0.90. The lowest congruence scores reported by McCrae et al. are 0.53 and 0.56 for Openness in Botswana and Nigeria, respectively. In the African and South Asian countries from Schmitt et al. (2007) , internal reliability for Extraversion, Agreeableness, and Conscientiousness is similar to what we report for the Tsimane. Because the samples from the developing countries in Schmitt et al. and McCrae et al. are primarily college students, more representative samples from these countries may have produced even lower congruence scores and internal reliability.

If the Big Five (or any other number of fixed traits) are not pan-human universals, then what could explain variability in personality structure? Nettle (2010) argued that personality items covary because they act synergistically. For example, he suggests that the fitness payoff to ambition is positive if sociability is also high; these traits thus covary as part of the Extraversion continuum. Similarly, the fitness payoff to imagination is positive if intellect is also high; thus, both traits covary along the Openness continuum. If the synergism of particular personality traits has different fitness consequences in different socioecological environments, we may not expect a universal structure of personality covariation. Behavioral genetic data support this possibility: Two independent dimensions of genetic variance are necessary to explain variation in each of the Big Five factors ( Jang, Livesley, Angleitner, Riemann, & Vernon, 2002 ). In different socioecologies, these independent genetic sources may not contribute to the same behavioral dispositions or experience parallel selection pressures ( Penke, Denissen, & Miller, 2007 ).

Variation in personality structure across populations need not derive from different patterns of covariation among genetic polymorphisms. Instead, different personality structures may arise from the facultative responses of individuals living in different socioecologies. In other words, individuals in different populations can share the same personality-relevant genetic architecture, but these genes may produce different effects in different environments. A growing body of work within behavioral ecology interprets personality variation as reaction norms that respond over ontogeny to individual condition and socioecological context ( Dingemanse, Kazem, Reale, & Wright, 2010 ; Sih, Bell, Johnson, & Ziemba, 2004 ). A working hypothesis is that coordinated traits might be facultatively calibrated based on cues underlying individual circumstances during development. The bundle of particular items and traits constituting human personality might act like conditional strategies ( Buss, 2009 ; Figueredo et al., 2011; Gangestad & Simpson, 2000 ; Lukaszewski & Roney, 2011 ; Nettle, 2010 ; Penke, 2010 ; Tooby & Cosmides, 1990 ). For example, men who are stronger and rated as more attractive are more likely to be extraverted, independent of a genetic polymorphism that also explains some of the covariance ( Lukaszewski & Roney, 2011 ). Variation in susceptibility to stress, which may underlie differences in neuroticism, has been linked to facultative calibration to stressors early in life ( Ellis, Jackson, & Boyce, 2006 ). It is an intriguing possibility that pan-human reaction norms shape not only intersocietal differences in average personality scores but also the structure of personality covariation itself, due to sustained socioecological differences across human populations. This hypothesis cannot be rejected in light of recent cross-cultural studies finding universal evidence of the Big Five, given the WEIRD-ness of most of the study populations. Indeed, any model of personality that specifies a fixed set of biologically based trait dimensions would be inconsistent with the results we report here. A comprehensive theory of personality would need to explain how particular conditions might lead to different combinations of calibrated and coordinated items, which then generate multidimensional personality structure, in varied socioecological settings and circumstances. Under a wide range of conditions, the FFM might adequately describe personality variation and necessarily so, but we still do not know why! We therefore speculate about some conditions that differ between WEIRD and small-scale subsistence societies in order to help explain our findings.

What features of Tsimane socioecology cause divergence from the Big Five pattern found in WEIRD populations? Individuals in all human societies face similar goals of learning important productive skills, avoiding environmental dangers, cooperating and competing effectively in social encounters, and finding suitable mates. In small-scale societies, however, individuals tend to live in small groups of closely related individuals with greatly reduced choice in social or sexual partners. There are also a limited number of niches by which cultural success may be measured, and proficiency may require abilities that connect items from different traits, thereby leading to low trait reliability and a trait structure other than the FFM. Among the Tsimane, success is defined largely in terms of ability to produce food and provision one’s family. Spouses rank each other primarily on these traits and are assortatively matched based on work effort ( Gurven et al., 2009 ). Leadership and allies outside of the extended family accrue to men who are outgoing, trustworthy, and generous among community members ( von Rueden et al., 2008 ). Women’s reputations are linked to similar traits and affect their ability to marshal intravillage exchange partnerships ( Rucas et al., 2006 ). Our industriousness and prosociality factors may reflect the different blends of traits conducive to success in the domestic versus the public sphere of Tsimane life. Furthermore, the orthogonality of these factors suggests their effects on fitness are partially independent. Lifetime reproductive success is higher for better producers ( Gurven & von Rueden, 2006 ) and for higher status individuals ( von Rueden, Gurven, & Kaplan, 2011 ), and status has a strong effect on reproduction even after controlling for productivity. It is possible that traits may vary more independently in WEIRD societies because of their greater niche diversity and specialization, whether in terms of professional careers or social groups. Success may require a coordinated assortment of fewer items that thereby bundle together in a larger number of factors.

Other considerations might also help explain our findings and would be important to test in other similar societies. Although extended families have relative political autonomy in many small-scale communities, an egalitarian ethic often curtails verbal expression of personal achievement ( Boehm, 1999 ). Thus, the costs and benefits of being extraverted may hinge on one’s level of agreeableness, which is suggested by the covariance of Extraversion and Agreeableness items in our prosociality factor. Indeed, Tsimane men whose voiced opinions are most influential in community meetings have more allies and are rated by their peers as more prosocial ( von Rueden et al., 2008 ). McCrae et al. (1998) and Cheung et al. (2001) argued that Extraversion and Agreeableness items have shown different factor structure in East Asian societies because they are more collectivist cultures in which interpersonal affiliation and obedience to authority are more normative. Small-scale societies such as the Tsimane can be characterized as collectivist only in terms of interpersonal affiliation: Their reliance on interhousehold exchange to buffer risk promotes consensual decision making and suppresses the emergence of formal authority ( Boehm, 1999 ; Cashdan, 1980 ).

Given the day-to-day risks of underproduction relative to subsistence needs, members of small-scale societies tend to be more risk averse ( Cancian, 1989 ; Cashdan, 1990 ; Kuznar, 2001 ), and new ideas, values, or experiences are typically met with conservatism. Furthermore, Tsimane and other small-scale populations in the tropics experience high levels of a variety of infectious pathogens ( Vasunilashorn et al., 2010 ), so a cautious and conservative approach to novel people, foods, and practices may reduce the risk of disease ( Schaller & Murray, 2008 ). In our EFAs, the Openness items of “original” and “ingenious” covary with socially desirable Extraversion and Agreeableness items; perhaps individuals who are the most interpersonally imbedded can best manage the risks of being open to new experiences. Items gauging artistic interest also covary with socially desirable traits; playing music and telling stories are the principal forms of artistic expression among the Tsimane and are most overt as “performance” in group settings. It is our impression that Tsimane who are more outgoing tend to be the most eager and creative singers and musicians. On the other hand, the Openness items of “curious” and “likes to reflect” positively covaried, respectively, with the Neuroticism item “moody” and the Agreeableness items “quarrelsome” and “rude.” This latter result supports our impression from the Tsimane and other small-scale societies that traits such as introspection and reflection are sometimes viewed as signs of depression or are viewed with suspicion. Openness exhibited low internal reliability and factor congruence in our study, similar to results from other developing countries (e.g., McCrae et al., 2005 ; Piedmont et al., 2002 ). Openness does not typically replicate in emic studies with Chinese subjects (e.g., Leung, Cheung, Zhang, Song, & Xie, 1997 ), suggesting collectivist norms may limit entrepreneurship and expression in ways that mimic the limited opportunities individuals face in small-scale societies. Resolution of these issues requires more studies of personality in non-WEIRD populations.

Whether the Big Five personality structure replicates in small-scale societies is crucial to claims of the universality of the FFM or any other fixed factor construct (e.g., HEXACO: Lee & Ashton, 2004 ; Big Two: Digman, 1997 ; General Factor of Personality: Musek, 2007 ). More important, data from small-scale societies contributes to our understanding of the evolution of human personality differences. The FFM and other structural approaches to personality variation are often criticized for a lack of theoretical justification (e.g., Block, 1995 ); conceptualizing personality dimensions as evolved motivational systems calibrated based on state-based cues from a particular socioecology has the potential to fill this void. Framing adaptive explanations of this sort in the study of human personality has a precedent in behavioral ecology. Concurrent developments in the biological sciences increasingly show that stable personalities, or “behavioral syndromes,” exist in many nonhuman species and can have substantial fitness consequences (for reviews of models and evidence, see Dingemanse & Wolf, 2010 ; Sih & Bell, 2008 ; Sih et al., 2004 ; Wolf & Weissing, 2010 ). The empirical study of Big Five traits and fitness outcomes in humans is still in its infancy (e.g., Alvergne et al., 2010 ; Eaves, Martin, Heath, Hewitt, & Neale, 1990 ; Nettle, 2005 ; Roberts, Kuncel, Shiner, Caspi, & Goldberg, 2007 ). Human personality research is therefore a ripe area for bridging theoretical models with rich empirical evidence ( Nettle & Penke, 2010 ). However, empirical evidence in humans must expand beyond the limited scope of WEIRD societies. What we can learn about personality variation in small-scale societies bears on arguments concerning the selection pressures responsible for shaping human personality traits and their structure. It is in small-scale societies that humans have lived for the majority of their existence; the socioecologies of ancestral hunter– gatherers and horticulturalists are the crucible that shaped much of human psychology and behavior. We therefore urge others to conduct similar studies of personality structure in other small-scale, indigenous societies.

We provide the first comprehensive test of the FFM in a small-scale, indigenous society—the Tsimane horticulturalists of Bolivia—and fail to robustly replicate the Big Five. We find significant covariance among items across the standard Big Five factors, based on two large samples of self- and spouse-reported personality. Tsimane personality variation may instead be organized along fewer and differently composed dimensions. We find evidence for a Tsimane Big Two organized according to prosociality and industriousness in the context of subsistence labor. Our current results require replication, with emic inventories and with other methods such as those based on behavioral observation or on peer reports by non-Tsimane. However, even if other methods were to reveal a Big Five structure, an explanation would still be needed for why verbal reports do not lead to the FFM among Tsimane, even after correction for response biases, but do almost everywhere else in the developed world.

Supplementary Material

Acknowledgments.

Funding was provided by the National Institutes of Health and the National Institute on Aging (Grants 2R01AG024119 and 2R56AG024119-06). We are grateful to the Tsimane for their hospitality and collaboration over the years. Gary Lewis provided helpful comments on a draft of this article. We also thank Aaron Lukaszewski for sharing ideas and commenting on a draft of the article.

Supplemental materials: http://dx.doi.org/10.1037/a0030841.supp

Contributor Information

Michael Gurven, Department of Anthropology, University of California, Santa Barbara.

Christopher von Rueden, Department of Anthropology, University of California, Santa Barbara.

Maxim Massenkoff, Department of Anthropology, University of California, Santa Barbara.

Hillard Kaplan, Department of Anthropology, University of New Mexico.

Marino Lero Vie, Tsimane Health and Life History Project, San Borja, Beni, Bolivia.

• Allik J, McCrae RR. Toward a geography of personality traits: Patterns of profiles across 36 culture. Journal of Cross-Cultural Psychology. 2004; 35 :13–28. [ Google Scholar ]
• Almagor M, Tellegen A, Waller NG. The Big Seven: A cross-cultural replication and further exploration of the basic dimensions of natural language of trait descriptions. Journal of Personality and Social Psychology. 1995; 69 :300–307. [ Google Scholar ]
• Alvergne A, Jokela M, Lummaa V. Personality and reproductive success in a high-fertility human population. Proceedings of the National Academy of Sciences, USA. 2010; 107 :11745–11750. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Ashton MC, Lee K, Goldberg LR, de Vries RE. Higher order factors of personality: Do they exist? Personality and Social Psychology Review. 2009; 13 :79–91. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Bäckström M, Björklund F, Larsson MR. Five-factor inventories have a major general factor related to social desirability which can be reduced by framing items neutrally. Journal of Research in Personality. 2009; 43 :335–344. [ Google Scholar ]
• Benet-Martínez V, John OP. Los Cinco Grandes across cultures and ethnic groups: Multitrait-multimethod analyses of the Big Five in Spanish and English. Journal of Personality and Social Psychology. 1998; 75 :729–750. [ PubMed ] [ Google Scholar ]
• Benet-Martínez V, Waller NG. Further evidence for the cross-cultural generality of the “Big Seven” model: Imported and indigenous Spanish personality constructs. Journal of Personality. 1997; 65 :567–598. [ Google Scholar ]
• Block J. A contrarian view of the five-factor approach to personality description. Psychological Bulletin. 1995; 117 :187–215. [ PubMed ] [ Google Scholar ]
• Boehm C. Hierarchy in the forest: The evolution of egalitarian behavior. Cambridge, MA: Harvard University Press; 1999. [ Google Scholar ]
• Bouchard TJ, Loehlin JC. Genes, evolution, and personality. Behavior Genetics. 2001; 31 :243–273. [ PubMed ] [ Google Scholar ]
• Buss DM. How can evolutionary psychology successfully explain personality and individual differences? Perspectives on Psychological Science. 2009; 4 :359–366. [ PubMed ] [ Google Scholar ]
• Cancian F. Economic behavior in peasant communities. In: Plattner S, editor. Economic anthropology. Stanford, CA: Stanford University Press; 1989. pp. 127–170. [ Google Scholar ]
• Caprara GV, Barbaranelli C, Borgogni L, Perugini M. The Big Five Questionnaire: A new questionnaire to assess the five-factor model. Personality and Individual Differences. 1993; 15 :281–288. [ Google Scholar ]
• Caprara GV, Perugini M. Personality described by adjectives: The generalizability of the Big Five to the Italian lexical context. European Journal of Personality. 1994; 8 :357–369. [ Google Scholar ]
• Cashdan EA. Egalitarianism among hunters and gatherers. American Anthropologist. 1980; 82 :116–120. [ Google Scholar ]
• Cashdan E. Introduction. In: Cashdan E, editor. Risk and uncertainty in tribal and peasant societies. Boulder, CO: Westview Press; 1990. pp. 1–16. [ Google Scholar ]
• Ceci SJ. How much does schooling influence general intelligence and its cognitive components? A reassessment of the evidence. Developmental Psychology. 1991; 27 :703–722. [ Google Scholar ]
• Cheung FM, Leung K. Indigenous personality measures: Chinese examples. Journal of Cross-Cultural Psychology. 1998; 29 :233–248. [ Google Scholar ]
• Cheung FM, Leung K, Zhang J-X, Sun H-F, Gan Y-Q, Song W-Z, Xie D. Indigenous Chinese personality constructs: Is the five-factor model complete? Journal of Cross-Cultural Psychology. 2001; 32 :407–433. [ Google Scholar ]
• Chicchon A. Chimane resource use and market involvement in the Beni Biosphere Reserve. Gainesville, Florida: University of Florida; 1992. Unpublished doctoral dissertation. [ Google Scholar ]
• Church AT, Lonner WJ. The cross-cultural perspective in the study of personality: Rationale and current research. Journal of Cross-Cultural Psychology. 1998; 29 :32–62. [ Google Scholar ]
• Costa PT, Jr, McCrae RR. Revised NEO Personality Inventory (NEO PI–R) and NEO Five-Factor Inventory (NEO-FFI): Professional manual. Odessa, FL: Psychological Assessment Resources; 1992. [ Google Scholar ]
• Costa PT, McCrae RR. C. Halverson, G. Kohnstamm, & R. Martin. The developing structure of temperament and personality from infancy to adulthood. Hillsdale, NJ: Erlbaum; 1994. 1994. Stability and change in personality from adolescence through adulthood; pp. 139–150. [ Google Scholar ]
• Denissen JJA, Penke L. Motivational individual reaction norms underlying the five-factor model of personality: First steps towards a theory-based conceptual framework. Journal of Research in Personality. 2008; 42 :1285–1302. [ Google Scholar ]
• De Raad B. An expedition in search of a fifth universal factor: Key issues in the lexical approach. European Journal of Personality. 1994; 8 :229–250. [ Google Scholar ]
• DeYoung CG. Higher-order factors of the Big Five in a multi-informant sample. Journal of Personality and Social Psychology. 2006; 91 :1138–1151. [ PubMed ] [ Google Scholar ]
• Di Blas L, Forzi M. An alternative taxonomic study of personality-descriptive adjectives in the Italian language. European Journal of Personality. 1998; 12 :75–101. [ Google Scholar ]
• Digman JM. Personality structure: Emergence of the five-factor model. Annual Review of Psychology. 1990; 41 :417–440. [ Google Scholar ]
• Digman JM. Higher order factors of the Big Five. Journal of Personality and Social Psychology. 1997; 73 :1246–1256. [ PubMed ] [ Google Scholar ]
• Dingemanse NJ, Kazem AJN, Reale D, Wright J. Behavioural reaction norms: Animal personality meets individual plasticity. Trends in Ecology & Evolution. 2010; 25 :81–89. [ PubMed ] [ Google Scholar ]
• Dingemanse NJ, Wolf M. Recent models for adaptive personality differences: A review. Philosophical Transactions of the Royal Society B: Biological Sciences. 2010; 365 :3947–3958. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Eaves LJ, Martin NG, Heath AC, Hewitt JK, Neale MC. Personality and reproductive fitness. Behavior Genetics. 1990; 20 :563–568. [ PubMed ] [ Google Scholar ]
• Ellis BJ, Jackson JJ, Boyce WT. The stress response systems: Universality and adaptive individual differences. Developmental Review. 2006; 26 :175–212. [ Google Scholar ]
• Figueredo AJ, Wolf PSA, Gladden PR, Olderbak S, Andrzejczak DJ, Jacobs WJ. Ecological approaches to personality. In: Buss DM, Hawley PH, editors. The evolution of personality and individual differences. Oxford, England: Oxford University Press; 2010. pp. 210–239. [ Google Scholar ]
• Gangestad SW, Simpson JA. The evolution of human mating: Trade-offs and strategic pluralism. Behavioral and Brain Sciences. 2000; 23 :573–587. [ PubMed ] [ Google Scholar ]
• Goldberg LR. An alternative “description of personality”: The Big-Five factor structure. Journal of Personality and Social Psychology. 1990; 59 :1216–1229. [ PubMed ] [ Google Scholar ]
• Guanzon-Lapeña MA, Church AT, Carlota AJ, Katigbak MS. Indigenous personality measures: Philippine examples. Journal of Cross-Cultural Psychology. 1998; 29 :249–270. [ Google Scholar ]
• Gurven M, Kaplan H, Zelada Supa A. Mortality experience of Tsimane Amerindians: Regional variation and temporal trends. American Journal of Human Biology. 2007; 19 :376–398. [ PubMed ] [ Google Scholar ]
• Gurven M, von Rueden C. Hunting, social status, and biological fitness. Biodemography and Social Biology. 2006; 53 :81–99. [ PubMed ] [ Google Scholar ]
• Gurven M, Winking J. Collective action in action: Pro-social behavior in and out of the laboratory. American Anthropologist. 2008; 110 :179–190. [ Google Scholar ]
• Gurven M, Winking J, Kaplan H, von Rueden C, McAllister L. A bioeconomic approach to marriage and the sexual division of labor. Human Nature. 2009; 20 :151–183. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Gurven M, Zanolini A, Schniter E. Culture sometimes matters: Intra-cultural variation in pro-social behavior among Tsimane Amerindians. Journal of Economic Behavior & Organization. 2008; 67 :587–607. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Henrich J, Heine SJ, Norenzayan A. The weirdest people in the world? Behavioral and Brain Sciences. 2010; 33 :61–83. [ PubMed ] [ Google Scholar ]
• Hofstee WKB, Ten Berge JMF, Hendriks AAJ. How to score questionnaires. Personality and Individual Differences. 1998; 25 :897–909. [ Google Scholar ]
• Huánca T. Tsimane’ indigeous knowledge, swidden fallow management, and conservation. University of Florida; 1999. Unpublished doctoral dissertation. [ Google Scholar ]
• Jang KL, Livesley W, Angleitner A, Riemann R, Vernon PA. Genetic and environmental influences on the covariance of facets defining the domains of the five-factor model of personality. Personality and Individual Differences. 2002; 33 :83–101. [ Google Scholar ]
• John OP. The “Big Five” factor taxonomy: Dimensions of personality in the natural language and in questionnaires. In: Pervin LA, editor. Handbook of personality: Theory and research. New York, NY: Guilford Press; 1990. pp. 66–100. [ Google Scholar ]
• King JE, Figueredo AJ. The five-factor model plus Dominance in chimpanzee personality. Journal of Research in Personality. 1997; 31 :257–271. [ Google Scholar ]
• Kuznar L. Risk sensitivity and value among Andean pastoralists: Measures, models, and empirical tests. Current Anthropology. 2001; 42 :432–440. [ Google Scholar ]
• Lee K, Ashton MC. The HEXACO Personality Inventory: A new measure of the major dimensions of personality. Multivariate Behavioral Research. 2004; 39 :329–358. [ PubMed ] [ Google Scholar ]
• Leung K, Cheung FM, Zhang J-X, Song W-Z, Xie D. The five-factor model of personality in China. In: Leung K, Kashima Y, Kim U, Yamaguchi S, editors. Progress in Asian social psychology. Vol. 1. Singapore: Wiley; 1997. pp. 231–244. [ Google Scholar ]
• Lukaszewski AW, Roney JR. The origins of extraversion: Joint effects of facultative calibration and genetic polymorphism. Personality and Social Psychology Bulletin. 2011; 37 :409–421. [ PubMed ] [ Google Scholar ]
• McCrae RR. Lonner WJ, Dinnel DL, Hayes SA, Sattler DN, editors. Cross-cultural research on the five-factor model of personality. 2002 Online readings in psychology and culture (Unit 6, Chapter 1). Retrieved from http://www.wwu.edu/~culture .
• McCrae RR, Costa PT., Jr Personality trait structure as a human universal. American Psychologist. 1997; 52 :509–516. [ PubMed ] [ Google Scholar ]
• McCrae RR, Costa PT, Jr, Del Pilar GH, Rolland JP, Parker WD. Cross-cultural assessment of the five-factor model: The revised NEO Personality Inventory. Journal of Cross-Cultural Psychology. 1998; 29 :171–188. [ Google Scholar ]
• McCrae RR, Terracciano A & 78 Members of the Personality Profiles of Cultures Project. Universal features of personality traits from the observer’s perspective: Data from 50 cultures. Journal of Personality and Social Psychology. 2005; 88 :547–561. [ PubMed ] [ Google Scholar ]
• McCrae RR, Yamagata S, Jang KL, Riemann R, Ando J, Ono Y, … Spinath FM. Substance and artifact in the higher-order factors of the Big Five. Journal of Personality and Social Psychology. 2008; 95 :442–455. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• McCrae RR, Zonderman AB, Costa PT, Jr, Bond MH, Paunonen SV. Evaluating replicability of factors in the Revised NEO Personality Inventory: Confirmatory factor analysis versus Procrustes rotation. Journal of Personality and Social Psychology. 1996; 70 :552–566. [ Google Scholar ]
• Musek J. A general factor of personality: Evidence for the Big One in the five-factor model. Journal of Research in Personality. 2007; 41 :1213–1233. [ Google Scholar ]
• Narayan S, Krosnick JA. Education moderates some response effects in attitude measurement. Public Opinion Quarterly. 1996; 60 :58–88. [ Google Scholar ]
• Nettle D. An evolutionary approach to the extraversion continuum. Evolution and Human Behavior. 2005; 26 :363–373. [ Google Scholar ]
• Nettle D. Evolutionary perspectives on the five-factor model of personality. In: Buss D, Hawley P, editors. The evolution of personality and individual differences. New York, NY: Oxford University Press; 2010. pp. 5–28. [ Google Scholar ]
• Nettle D, Penke L. Personality: Bridging the literatures from psychology and behavioural ecology. Philosophical Transactions of the Royal Society B: Biological Sciences. 2010; 365 :4043–4050. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Ortiz FA, Church AT, Vargas-Flores JDJ, Ibáñez-Reyes J, Flores-Galaz M, Iuit-Briceño JI, Escamilla JM. Are indigenous personality dimensions culture-specific? Mexican inventories and the five-factor model. Journal of Research in Personality. 2007; 41 :618–649. [ Google Scholar ]
• Paunonen SV, Ashton MC. The structured assessment of personality across cultures. Journal of Cross-Cultural Psychology. 1998; 29 :150–170. [ Google Scholar ]
• Paunonen SV, Ashton MC, Jackson DN. Nonverbal assessment of the Big Five personality factors. European Journal of Personality. 2001; 15 :3–18. [ Google Scholar ]
• Peabody D, De Raad B. The substantive nature of psycholexical personality factors: A comparison across languages. Journal of Personality and Social Psychology. 2002; 83 :983–997. [ PubMed ] [ Google Scholar ]
• Penke L. Bridging the gap between modern evolutionary psychology and the study of individual differences. In: Buss DM, Hawley PH, editors. The evolution of personality and individual differences. Oxford, England: Oxford University Press; 2010. pp. 243–279. [ Google Scholar ]
• Penke L, Denissen JJA, Miller GF. The evolutionary genetics of personality. European Journal of Personality. 2007; 21 :549–587. [ Google Scholar ]
• Perugini M, Leone L. Construction and validation of a Short Adjectives Checklist to measure Big Five (SACBIF) European Journal of Psychological Assessment. 1996; 12 :33–42. [ Google Scholar ]
• Piedmont RL, Bain E, McCrae RR, Costa PT., Jr . The applicability of the five-factor model in a sub-Saharan culture: The NEO-PI–R in Shona. In: McCrae RR, Allik J, editors. The five-factor model of personality across cultures. New York, NY: Kluwer Academic; 2002. pp. 155–173. [ Google Scholar ]
• Ramírez-Esparza N, Gosling SD, Benet-Martínez V, Potter JP, Pennebaker JW. Do bilinguals have two personalities? A special case of cultural frame switching. Journal of Research in Personality. 2006; 40 :99–120. [ Google Scholar ]
• Rammstedt B, Goldberg LR, Borg I. The measurement equivalence of Big Five factor markers for persons with different levels of education. Journal of Research in Personality. 2010; 44 :53–61. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Riemann R, Angleitner A, Strelau J. Genetic and environmental influences on personality: A study of twins reared together using the self- and peer-report NEO FFI scales. Journal of Personality. 1997; 65 :449–475. [ Google Scholar ]
• Roberts BW, Kuncel NR, Shiner R, Caspi A, Goldberg LR. The power of personality: The comparative validity of personality traits, socio-economic status, and cognitive ability for predicting important life outcomes. Perspectives on Psychological Science. 2007; 2 :313–345. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Rolland JP. Cross-cultural generalizability of the five-factor model of personality. In: McCrae RR, Allik J, editors. The five-factor model of personality across cultures. New York, NY: Kluwer Academic; 2002. pp. 7–28. [ Google Scholar ]
• Rucas S, Gurven M, Kaplan H, Winking J, Gangestad S, Crespo M. Female intrasexual competition and reputational effects on attractiveness among the Tsimane of Bolivia. Evolution and Human Behavior. 2006; 27 :40–52. [ Google Scholar ]
• Schaller M, Murray DR. Pathogens, personality, and culture: Disease prevalence predicts worldwide variability in sociosexuality, extraversion, and openness to experience. Journal of Personality and Social Psychology. 2008; 95 :212–221. [ PubMed ] [ Google Scholar ]
• Schmitt DP, Allik J, McCrae RR, Benet-Martínez V, Alcalay L, Ault L. The geographic distribution of Big Five personality traits: Patterns and profiles of human self description across 56 nations. Journal of Cross-Cultural Psychology. 2007; 38 :173–212. [ Google Scholar ]
• Schniter E. Why old age? Non-material contributions and patterns of aging among older adult Tsimane. Santa Barbara, California: University of California; 2009. Unpublished doctoral dissertation. [ Google Scholar ]
• Sih A, Bell AM. Insights for behavioral ecology from behavioral syndromes. Advances in the Study of Behavior. 2008; 38 :227–281. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Sih A, Bell AM, Johnson JC, Ziemba RE. Behavioural syndromes: An integrative overview. Quarterly Review of Biology. 2004; 79 :241–277. [ PubMed ] [ Google Scholar ]
• Soto CJ, John OP, Gosling SD, Potter J. The development psychometrics of the Big Five self-reports: Acquiescence, factor structure, coherence, and differentiation from ages 10 to 20. Journal of Personality and Social Psychology. 2008; 94 :718–737. [ PubMed ] [ Google Scholar ]
• Szirmák Z, De Raad B. Taxonomy and structure of Hungarian personality traits. European Journal of Personality. 1994; 8 :95–117. [ Google Scholar ]
• Tabachnick BG, Fidell LS. Using multivariate statistics. Boston, MA: Allyn & Bacon; 2001. [ Google Scholar ]
• Thompson ER. Development and validation of an international English Big-Five mini markers. Personality and Individual Differences. 2008; 45 :542–548. [ Google Scholar ]
• Tooby J, Cosmides L. On the universality of human nature and the uniqueness of the individual: The role of genetics and adaptation. Journal of Personality. 1990; 58 :17–67. [ PubMed ] [ Google Scholar ]
• Triandis H. Cross-cultural perspectives on personality. In: Hogan R, Johnson J, Briggs S, editors. Handbook of personality psychology. San Diego, CA: Academic Press; 1997. pp. 439–464. [ Google Scholar ]
• Vasunilashorn S, Crimmins EM, Kim JK, Winking J, Gurven M, Kaplan H, Finch CE. Blood lipids, infection, and inflammatory markers in the Tsimane of Bolivia. American Journal of Human Biology. 2010; 22 :731–740. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• von Rueden C, Gurven M, Kaplan H. Multiple dimensions of male social status in an Amazonian society. Evolution and Human Behavior. 2008; 29 :402–415. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• von Rueden C, Gurven M, Kaplan H. Why do men seek status? Fitness payoffs to dominance and prestige. Proceedings of the Royal Society B: Biological Sciences. 2011; 278 :2223–2232. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Wiggins JS, Trapnell PD. Personality structure: The return of the Big Five. In: Hogan R, Johnson J, Briggs S, editors. Handbook of personality psychology. San Diego, CA: Academic Press; 1997. pp. 737–765. [ Google Scholar ]
• Wolf M, Weissing FJ. An explanatory framework for adaptive personality differences. Philosophical Transactions of the Royal Society B: Biological Sciences. 2010; 365 :3959–3968. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Yamagata S, Suzuki A, Ando J, Ono Y, Kijima N, Yoshimura K, … Jang KL. Is the genetic structure of human personality universal? A cross-cultural twin study from North America, Europe, and Asia. Journal of Personality and Social Psychology. 2006; 90 :987–998. [ PubMed ] [ Google Scholar ]

Academia.edu no longer supports Internet Explorer.

To browse Academia.edu and the wider internet faster and more securely, please take a few seconds to  upgrade your browser .

•  We're Hiring!
•  Help Center

Big Five Personality Traits

• Most Cited Papers
• Most Downloaded Papers
• Newest Papers
• Last »
• THE BIG FIVE MODEL Follow Following
• Five Factor Model Follow Following
• The big five Follow Following
• Structure Of Personality Follow Following
• Time Perspective Follow Following
• Comparative psychology Follow Following
• Marketing Psychology Follow Following
• Test anxiety Follow Following
• Electricity Market Follow Following
• Anger Follow Following

Enter the email address you signed up with and we'll email you a reset link.

• Academia.edu Journals
•   We're Hiring!
•   Help Center
• Find new research papers in:
• Health Sciences
• Earth Sciences
• Cognitive Science
• Mathematics
• Computer Science
• Academia ©2024

• DOI: 10.24036/jea.v6i3.1624
• Corpus ID: 271781030

Pengaruh Big Five Personality terhadap Skeptisisme Profesional Auditor

• Laura Callista Wijaya , Charoline Cheisviyanny
• Published in JURNAL EKSPLORASI AKUNTANSI 3 August 2024
• Business, Psychology

56 References

Pengaruh big five personality pada kinerja auditor kantor akuntan publik provinsi bali.

• Highly Influential

Big Five Personality Traits dan Skeptisisme Profesional Auditor

Pengaruh extraversion dan neuroticism terhadap audit judgment, pengaruh faktor-faktor personal terhadap skeptisisme profesional auditor, the role of anticipatory socialization as a mediating variable between the big five personality traits and professional skepticism, the relationship between accounting students’ personality, professional skepticism and anticipatory socialization, the mediating role of independence in the relationship between auditors’ conscientiousness & openness and professional skepticism, pengaruh gender, tipe kepribadian, tingkat pendidikan, dan pengalaman audit terhadap skeptisme auditor internal pemerintah (studi empiris pada inspektorat provinsi jawa timur), personality tests in accounting research, examining the cheats: the role of conscientiousness and excitement seeking in academic dishonesty, related papers.

Showing 1 through 3 of 0 Related Papers