U.S. flag

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

Logo of springeropen

Risk Factors for School Absenteeism and Dropout: A Meta-Analytic Review

Jeanne gubbels.

Research Institute of Child Development and Education, University of Amsterdam, P.O. Box 15780, Nieuwe Achtergracht 127, 1018 WS Amsterdam, The Netherlands

Claudia E. van der Put

Mark assink.

School absenteeism and dropout are associated with many different life-course problems. To reduce the risk for these problems it is important to gain insight into risk factors for both school absenteeism and permanent school dropout. Until now, no quantitative overview of these risk factors and their effects was available. Therefore, this study was aimed at synthesizing the available evidence on risk factors for school absenteeism and dropout. In total, 75 studies were included that reported on 781 potential risk factors for school absenteeism and 635 potential risk factors for dropout. The risk factors were classified into 44 risk domains for school absenteeism and 42 risk domains for dropout. The results of a series of three-level meta-analyses yielded a significant mean effect for 28 school absenteeism risk domains and 23 dropout risk domains. For school absenteeism, 12 risk domains were found with large effects, including having a negative attitude towards school, substance abuse, externalizing and internalizing problems of the juvenile, and a low parent-school involvement. For dropout, the risk domains having a history of grade retention, having a low IQ or experiencing learning difficulties, and a low academic achievement showed large effects. The findings of the current study contribute to the fundamental knowledge of the etiology of school absenteeism and dropout which in turn contributes to a better understanding of the problematic development of adolescents. Further, more insight into the strength of effects of risk factors on school absenteeism and dropout is important for the development and improvement of both assessment, prevention and intervention strategies.

Introduction

Problematic school absenteeism is associated with many different life-course problems, such as risky sexual behavior, teenage pregnancy, psychiatric disorders, externalizing behavior, delinquency, and the abuse of alcohol, tobacco, marijuana, and other substances (see, for example, Chou et al. 2006 ; Egger et al. 2003 ; Jaafar et al. 2013 ). In addition, youth showing excessive absenteeism are at high risk for permanent dropout from school (Kearney 2008a ), which may lead to economic deprivation and different mental, social, occupational, and marital problems in adulthood (Kogan et al. 2005 ; Tramontina et al. 2001 ). To reduce the risk for these problems, it is important to gain insight into risk factors for both problematic school absenteeism (i.e., temporary periods of unexcused school absence) and permanent school dropout. School absenteeism in youth refers to excused or unexcused absences from elementary or secondary (middle/high) school (Kearney 2008a ). Whereas excused absenteeism (e.g., absences related to medical illness or injury) could be viewed as non-problematic, unexcused and excessive absenteeism is a problem of serious concern that affects many school systems around the world. Absenteeism rates differ depending on the definition and measurement period. According to the National Center for Education Statistics ( 2018 ), 13% of the 8 th graders, 14% of the 10 th graders, and 15% of the 12 th graders were absent at least three days a month, and 6, 5, and 6% were absent at least five days a month, respectively. Until now, many studies on risk factors for school absenteeism and dropout have been performed, but no clear overview of risk factors and their effects was available. The aim of the present study was to provide such an overview by statistically summarizing effects of risk factors by conducting a series of meta-analyses.

Problematic school absenteeism (from now on referred to as school absenteeism) does not refer to a single concept, but to various concepts, including school refusal (absenteeism due to the child’s emotional distress, especially anxiety and depression; King and Bernstein 2001 ), school phobia (fear-based absenteeism; Tyrrell 2005 ), truancy (unexcused, illegal, non-anxiety-based absenteeism, which is often linked to a lack of parental monitoring, delinquency, academic problems, or social conditions such as homelessness or poverty; Fremont 2003 ) and absence from specific lessons. In their interdisciplinary model of school absenteeism, Kearney ( 2008a ) argue that these concepts of school absenteeism are influenced by multiple child, parent, family, peer, school, and community factors. They argue that school absenteeism cases are caused by multiple factors and that the key influential factors are interrelated (e.g., child and parent psychopathology). They also argue that school absenteeism can deteriorate over time from acute, but relatively harmless and occasional absenteeism into regular, and even permanent absenteeism in the form of dropping out of school. This view on how school absenteeism and dropout evolve is in line with the ecological perspective on child development of Bronfenbrenner ( 1979 , 1986 ). In his influential ecological model, Bronfenbrenner noted that the child interacts with different social ecological systems surrounding the child, such as the family, peers, and the school environment (microsystem), the extended family (exosystem), and the culture, laws, and social-political conditions (macrosystem). In each of these systems, risk factors can be present that increase the risk of negative child behavior, of which school absenteeism is an example. Bronfenbrenner assumed that risk factors in more proximal social systems exert more influence on the child’s development and behavior than risk factors in more distal social systems. Therefore, primary studies aimed at determining risk factors for school absenteeism and school dropout are mainly focused on child-related factors and factors present in the microsystems directly surrounding the child, such as family-, peer-, and school-related factors.

In theoretical models for explaining school absenteeism and dropout such as described above, risk factors play a critical role. Therefore, a large body of research has been directed on identifying risk factors for school absenteeism and school dropout. Some of these risk factors are related to characteristics of the child (e,g., the child’s age [the risk for school absenteeism increases as children become older], internalizing problems, externalizing problems, and a poor physical health), characteristics of the parent (e.g., parental psychiatric problems and parental unemployment), characteristics of the family (e.g. a low socio-economic status and family break-up), characteristics of the school (e.g. large classes, high retention rates, and a poor quality of teachers) or characteristics of the peer group (e.g. antisocial, truant, or delinquent peers). Primary studies examining risk factors for school absenteeism and dropout often show a wide variation in effect size magnitude. Previous reviews of these studies have provided an overview of risk factors or potential causes for school absenteeism (and related concepts) and dropout. Kearney ( 2008b ), for example, reviewed contemporary research on, among other things, the contextual risk factors for school absenteeism and school refusal behavior. Furthermore, Berends and Van Diest ( 2014 ) summarized the protective and risk factors for school absenteeism, and King and Bernstein ( 2001 ) reviewed studies on problematic family functioning as an important factor contributing to school refusal. However, these reviews were merely qualitative in nature, and until today, the literature on risk factors for school absenteeism and dropout has never been meta-analytically or quantitatively synthesized. In a meta-analysis, the divergent findings of studies on (effects of) risk factors can be summarized to increase insight into whether or not a factor should be designated as a risk factor, and what the true effect of a particular risk factor is. Accordingly, more insight can be gained into all risk factors that play a role in school absenteeism and dropout, leading to a better understanding of the etiology of these problems.

An overview of the variables that are true risk factors for school absenteeism and dropout is also relevant for clinical practice, as this may contribute to the development or improvement of instruments for risk and needs assessment. Risk assessment instruments assess which static (unchangeable in treatment) and dynamic (changeable in treatment) risk factors are present in the environment of a child, and are needed in determining which children should be offered an (preventive) intervention, and with what intensity these children should treated. Needs assessment instruments assess only dynamic risk factors (i.e. the care needs), and are needed in order determining what factors should be targeted in an intervention, so that the risk for school absenteeism or dropout is reduced. Both type of instruments originate from the risk and need principle of the Risk Need Responsivity (RNR) model (Andrews and Bonta 2010 ; Andrews et al. 1990 ). This model is used in judicial care as a guidance for offering effective offender assessment and treatment services, and its effectiveness has been proved in several review studies (see, for instance, Andrews et al. 1990 ; Andrews and Dowden 1999 ). It can be assumed that this model also applies to problematic and chronic school absenteeism, since criminal recidivism, school absenteeism, and school dropout can all be explained by an accumulation of risk factors in different domains. In addition, there is an overlap between risk factors for school absenteeism and delinquency (Van der Woude et al. 2017 ).

The present study, then, is important for several reasons. First, examining the effects of different risk factors for school absenteeism and dropout increases the fundamental knowledge of the etiology of these behavioral problems. Second, more insight into the effects of risk factors contributes to the development or improvement of risk and needs assessment instruments. Currently, there are hardly any risk and needs assessment instruments available that assess all relevant risk factors for school absenteeism and dropout, even though such instruments are required for properly referring at-risk juveniles to the most appropriate interventions for reducing risks. Third, the results of this study can support the development and improvement of interventions aimed at preventing (new occurrences of) school absenteeism or dropout. Information on the magnitude of dynamic risk factor effects is essential for determining which risk factors can best be addressed in these interventions.

The Current Study

This study aimed to synthesize the available evidence on risk factors for school absenteeism and dropout. Specifically, this study was guided by the research questions (1)“What factors can be designated as risk factors for school absenteeism and what is their impact?” and (2) “What factors can be designated as risk factors for school dropout and what is their impact?”. In answering these questions, each (potential) risk factor that was examined in a primary studies was classified into a risk domain, which is as a (broad) group of risk factors that are similar in nature. Next, an overall mean effect was estimated for each of these risk domains in a separate meta-analysis. Finally, as previous literature showed large gender differences in motives for school absenteeism and school dropout (e.g., De Baat and Foolen 2012 ; Teasley 2004 ), it was assumed that (effects of) risk factors do not need to be equal for boys and girls. Therefore, this study aimed to answer the following additional research question: (3) “How are risk factor effects influenced by gender?”. To address this final question, the percentage of boys in primary study samples was tested as moderator of the overall effect of each risk domain.

Inclusion and Exclusion Criteria

To select relevant studies, several inclusion and exclusion criteria were formulated. First, studies had to examine the effect of at least one (potential) risk factor for school absenteeism and/or dropout. In the current meta-analysis, school absenteeism refers to problematic school absenteeism, which was defined as unexcused absences from school (Kearney 2008a ). As described in the Introduction, problematic school absenteeism refers to various concepts, including missing or skipping classes, school non-attendance, and school refusal. Therefore, primary studies reporting on problematic school absenteeism and/or on one or more of these individual concepts were all included. Studies reporting on permitted or excused school absence were not included. School dropout was defined as leaving school prior to earning a high school credential (Kearney 2008b ).

Second, only studies examining school absenteeism and/or dropout in primary schools (kindergarten and elementary schools) and secondary schools (middle schools, junior high schools, and high schools) were included. Studies examining absence from college or other forms of post-secondary education were excluded.

Third, as risk factors must precede an outcome (Kraemer et al. 1997 ), only effect sizes of (potential) risk factors that were present prior to the school absenteeism or school dropout were included. Specifically, primary studies had to report on at least one association between school absenteeism or school dropout and a factor preceding these events, or a factor of which reasonably could be assumed to precede the absenteeism or school dropout based on information described in the primary study. Studies with a longitudinal research design (in which subjects were followed over time) as well as cross-sectional studies (in which subjects were examined at a single point in time) were included. However, factors reported in cross-sectional studies were only included if the factors were already present prior to any (potential) school absenteeism or dropout. This third criterion was to ensure that antecedents of school absenteeism were examined instead of consequences.

Fourth, studies had to report on (1) a measure of bivariate association between a factor and school absenteeism or dropout (e.g., a correlation coefficient) or (2) sufficient information for calculating such an association.

Fifth, given that risk factors for school absenteeism and dropout may be very different in prevalence and nature across cultural settings, only studies that were performed in Western countries were included (i.e., European countries, Australia, New Zealand, Canada, and the US). All primary studies had to be written in Dutch and English to be included.

Sixth, only studies published in peer-reviewed scientific journals or dissertations accessible to the authors of this review were included. Published studies have survived some form of a refereeing and editing process (Dunkin 1996 ), and although dissertations are not peer-reviewed, they have been evaluated by supervising committees and therefore controlled for quality at least to some extent. As this is not the case for unpublished studies, and as unpublished studies are far more difficult to locate, only published studies and dissertations were searched for and included.

Finally, the aim was not to perform a meta-analysis of the effects of treatment or preventive strategies for reducing school absenteeism and dropout, and because treatment effects may influence risk factor effects, no effects of potential risk factors that are reported in studies examining treatment effects were extracted.

Search Strategy

Until May 2019, multiple electronic databases were searched to identify relevant studies: Google, Google Scholar, ScienceDirect, PsycINFO, Web of Science, and Sociological Abstracts. The following keywords were used: “truan*”, “dropout”, “drop-out”, “school attendan*”, “school non-attendan*”, “school disengage*”, “class-cutting”, “school refus*”, “school absent*”, “risk factor*”, and “correlate*” (the asterisk represents one or more wildcard characters). Keywords related to “risk factors” were combined with keywords related to “school absenteeism” or “dropout”. Further, the reference list of several relevant reviews and reports were screened (e.g., Berends and Van Diest 2014 ; De Baat and Foolen 2012 ; Hammond et al. 2007 ; Kearney 2008b ; Teasley 2004 ) for relevant studies. Finally, the reference sections of the included primary studies were screened.

These search methods resulted in 4618 studies. After deduplication and the exclusion of studies based on their title or abstract, 220 studies remained of which the full text was evaluated. Finally, 75 studies met all inclusion criteria and were included in the current study. These studies reported on 71 independent samples. Figure ​ Figure1 1 presents a flow chart of the search of studies and Table ​ Table1 1 presents the characteristics of the included studies.

An external file that holds a picture, illustration, etc.
Object name is 10964_2019_1072_Fig1_HTML.jpg

Flow chart of search results

Characteristics of included studies

Pub. year publication year, Focus the focus of the study (school absenteeism or dropout), N total sample size, n (absent) number of school absentees or dropouts in the sample, n (non-absent) number of non-absentees or non-dropouts in the sample, % Boys percentage of boys in the sample, # Risk factors total number of risk factors that were retrieved from the study, # Child factors number of child factors that were retrieved from the study, # Family factors number of family factors that were retrieved from the study, # School factors number of school factors that were retrieved from the study, # Peer factors number of peer factors that were retrieved from the study, (2) A primary study using the same sample as the study mentioned directly above, Absenteeism/Dropout both school absenteeism and dropout were assessed in study

a If more than one sample size was reported in a study (varied by effect size), the range was displayed

b If more than one number of truants or dropouts was reported in a study (varied by effect size), the range was displayed

c If more than one number of non-truants or non-dropouts was reported in a study (varied by effect size), the range was displayed

d If more than one percentage of boys was reported in a study (due to multiple included samples; varied by effect size), all the percentages were displayed, separated by a ‘/’

Study Coding

Following the guidelines proposed by Lipsey and Wilson ( 2001 ), a coding form was developed to code all included primary studies. The primary interest was to synthesize all effects of risk factors that were similar in nature. Across all effect sizes that could be extracted from all included studies, there were too many risk factors to examine individually. For valid and intelligible analyses, each individual risk factor was classified into risk domains, which can be defined as categories of risk factors that are (more or less) similar in nature. According to the interdisciplinary model of school absenteeism of Kearney ( 2008a ) and the ecological model of Bronfenbrenner ( 1979 ; 1986 ; see Introduction), these risk domains are related to (1) characteristics of the child; (2) characteristics of parents or caretakers, and the family; (3) characteristics of the school; or (4) characteristics of peer relationships and interactions with peers. For each extracted risk factor, it was first determined whether the factor was related to the child, the family, the school, or the peers. Next, a risk factor was further classified into more specific risk domains, and this procedure was done separately for school absenteeism and dropout. The online Appendix A shows an overview of the risk domains that were examined in this review. In the end, all risk factors for school absenteeism were classified into one of 44 mutually exclusive risk domains, of which 24 were related to child characteristics, 11 to family characteristics, 6 to school characteristics, and 3 to peer and peer-interaction characteristics. Risk factors for school dropout were classified into 42 mutually exclusive risk domains, of which 23 were related to child characteristics, 12 to family characteristics, 4 to school characteristics, and 3 to peer and peer interaction characteristics.

For descriptive purposes, several sample and study characteristics were coded. However, it was decided to only examine the moderating effect of one sample characteristic, namely the percentage of boys within the sample. This variable was tested as a moderator, as it is known that there can be large gender differences in motives for school absenteeism and dropout (e.g., De Baat and Foolen 2012 ; Teasley 2004 ). In coding studies for meta-analytic research, it is common practice to retrieve a large amount of information from primary studies (see for instance, Cooper 2010 ; Lipsey and Wilson 2001 ), after which the moderating effect of a variety of study, sample, and research design descriptors is tested. However, since the problem of multiple testing often dealt with in primary studies (e.g., Tabachnik and Fidell 2013 ) is equally present in meta-analytic research, it was decided to only test the variable that seemed most relevant in light of the aims of the present review. Further, in order to gain sufficient statistical power in the moderator analyses, the variable percentage of boys in the sample was only tested as a moderator when this variable was based on at least five studies. The other coded variables did not meet this criterion, which was also reason not to test any other variable as a moderator within the risk domains.

In coding all included studies, two coding rounds were completed. First, 10 studies that were eligible for inclusion (7 school absenteeism studies and 3 dropout studies, reporting on a total of 282 risk factors) were randomly selected and coded by the first author and an and an independent assistant researcher. Next, the independent codings were compared and percentages of agreement were calculated. A perfect agreement (100%) was found for the percentage of boys in the sample, and the number of extracted effect sizes from each primary study. The agreement for the double-coded effect sizes was calculated by dividing the number of matching codings (268) by the total number of double-coded effect sizes (282), which was 95%. All discrepancies in the 5% non-matching effect size codings were discussed by the two coders until full consensus was reached. In the second coding round, the first author coded the remaining 65 studies. Finally, the classification of every extracted (potential) risk factor into risk domains was discussed by the first, second, and third author of this study. Therefore, the interrater agreement for the risk domain variable was perfect (100%).

Calculation of Effect Sizes and Statistical Analyses

In this review, the correlation coefficient ( r ) was chosen as common effect size for risk factor effects, meaning that a correlation was calculated for each extracted (potential) risk factor. The correlations were directly obtained from the included studies, or calculated using information that was reported in the studies (such as proportions, means and standard deviations, odds-ratio’s, or F or t values). In these calculations, the formulas of Ferguson ( 1966 ), Rosenthal ( 1994 ), and Lipsey and Wilson ( 2001 ) were used. A positive r value was assigned to a factor that was more present in youth showing school absenteeism or dropout than in youth not showing these problems, whereas a negative r value was assigned to a factor that was less present in youth showing school absenteeism or dropout. If a risk factor effect was reported as non-significant in primary studies without further statistical information to calculate the actual effect size, an effect size of zero was assigned to the factor (see also Durlak and Lipsey 1991 ). This procedure was applied to one study, in which two factors were described as non-significant. After all correlation coefficients were obtained, the r values were transformed into Fisher's z values, as correlations are non-normally distributed (see, for instance, Lipsey and Wilson 2001 ).

Because most studies reported on more than one risk factor for school absenteeism or dropout, a traditional random effects (two-level) model was extended to a three-level random effects model (Cheung 2014 ; Houben et al. 2015 ; Van den Noortgate et al. 2013 , 2014 ). A major advantage of this three-level approach to meta-analysis is that all relevant effects reported in each primary study can be included, implying that all relevant information is preserved. As a result, there no information is lost and (moderator) effects can be estimated more precisely and with maximum power in the statistical analyses (Assink and Wibbelink 2016 ). In a three-level random effects meta-analytic model, three sources of variance are taken into account: sampling variance of the observed effect sizes (Level 1), variance between effect sizes extracted from the same study (Level 2), and variance between studies (Level 3). In an intercept-only model, the intercept represents the estimate of the overall or mean effect of a single risk domain. If variation in effect sizes extracted from the same study (i.e., level 2 variance) and/or variation in effect sizes extracted from different studies (i.e., level 3 variance) was significant, the model was extended with the potential moderating variable percentage of boys to determine whether this variable can explain any significant variance. In a number of included studies, variables were examined as risk factors using the same sample. As this induces dependency in effect sizes that are extracted from these studies, the same study identification number was given to these studies, so that effect size dependency is accounted for.

In the statistical environment R (version 3.5.1; R Core Team 2015 ), the function “rma.mv” of the metafor-package (Viechtbauer 2010 ) was used to conduct the statistical analyses. The R syntaxes were written so that the three sources of variance were modeled (Assink and Wibbelink 2016 ). In testing individual regression coefficients and calculating corresponding confidence intervals, a t -distribution was used (Knapp and Hartung 2003 ). To determine the significance of the level 2 and level 3 variance, the full model was compared to a model excluding one of these variance parameters in two separate log-likelihood ratio tests. If significant level-2 and/or level-3 variance was detected, the distribution of effect sizes was considered to be heterogeneous. This indicated that effect sizes could not be treated as estimates of one common effect size, meaning that moderator analyses could be performed. All model parameters were estimated using the restricted maximum likelihood estimation method. Prior to the analyses, a dichotomous dummy variable was created for each category of a discrete variable and continuous variables were centered around their mean. The log-likelihood-ratio-tests were performed one-tailed and all other tests were performed two-tailed. A p -value < 0.05 was considered as statistically significant. Finally, it should be noted that all significant and non-significant results of all performed analyses are reported. No significant or non-significant result of any analysis was left out.

Assessment of Bias

Despite an extensive search for studies on risk factors for school absenteeism and dropout, it is possible that relevant studies were missed due to limitations in the search strategy or different forms of bias, such as publication bias or subjective reporting bias. To examine whether (a form of) bias was present in the estimated overall effects of risk domains, three analyses were conducted that are all three based on the association between effect size and sample size that is expected when bias is present in the effect sizes that are to be synthesized. First, a funnel-plot-based trim and fill method was conducted (Duval and Tweedie 2000a , 2000b ). This means that in case of an asymmetrical distribution of effect sizes (i.e., an asymmetrical funnel plot), the symmetry of the distribution is restored by imputing effect size estimates from “missing” studies. Effect sizes imputed to the left of the estimated mean effect imply that below average effect sizes were underrepresented and that the estimated mean effect may be an overestimation of the true effect. On the other hand, imputation of effect sizes to the right of the estimated mean effect indicates that above average effect sizes were underrepresented and that the estimated mean effect may be an underestimation of the true effect. Second, a three-level funnel plot test was conducted in which effect sizes were regressed on the sample sizes in a 3-level meta-analytic model, in which effect size dependency is accounted for. In this model, a significant slope is an indication of bias. Third, an adapted Egger”s test was conducted in which effect sizes were regressed on standard errors in a 3-level meta-analytic model. In this test, effect size dependency was also accounted for and a significant slope is once again an indication of bias. These bias assessment analyses were also performed in the R environment (Version 3.5.1; R Core Team 2015 ) with the functions “trimfill” and “rma.mv” of the metafor package (Viechtbauer 2010 ).

In total, k  = 75 studies published between 1978 and 2019 were included with k =  43 studies reporting on factors for school absenteeism and k =  33 studies reporting on factors for dropout. For specifically school absenteeism, 43 studies with 41 non-overlapping samples ( N  = 243,296 pupils) were included, from which 781 effect sizes were extracted. The average percentage of boys in the samples of these studies was 47.9%. All included studies together reported on at least n  = 26,230 absentees and at least n  = 189,437 non-absentees. Exact numbers of these groups could not be given, as in some studies the specific number of absentees and non-absentees was not reported. The included studies were conducted in the USA ( k  = 21), Canada ( k  = 3), Australia ( k  = 1), and Europe ( k  = 16).

The 33 studies on school dropout used 31 non-overlapping samples with a total sample size ( N ) of 136,392 pupils. These studies examined at least n  = 21,625 school dropouts and at least n  = 95,813 non-dropout (again, some of the dropout studies did not report on the specific number of dropouts and non-dropouts), and reported on 635 effect sizes. The average percentage of boys in the samples of these studies was 51.8%. The dropout studies were conducted in USA ( k  = 21), Canada ( k  = 5), and Europe ( k  = 6).

Overall Effects of Risk Domains for School Absenteeism

Table ​ Table2 2 presents an estimated overall effect for each of the 44 risk domains for school absenteeism in descending order, separately for child-, family-, school- and peer related risk domains. The overall effects of 28 domains were significant and positive in direction (including 16 child-related risk domains, 9 family-related risk domains, and 3 school-related risk domains), implying that these domains can be regarded as true risk domains for school absenteeism. The magnitude of the effects of these risk domains ranged from small (i.e., r =  0.099 for “low IQ/learning difficulties”) to large (i.e., r =  0.553 for “having a negative school attitude”) based on the criteria of Rice and Harris ( 2005 ) for interpreting effect sizes. Significant large overall effects ( r  ≥ 0.252) were found for 11 risk domains (indicated in Table ​ Table2 2 with “ a ”), including the child related risk domains “having a negative school attitude”, “anti-social behavior/cognitions”, “smoking”, “drug abuse”, “alcohol abuse”, “other internalizing problems”, “psychiatric symptoms or disorders”, and “being a sexual minority”; the family related risk domains “low parental school involvement” and “history of child abuse victimization”; and the school risk domain “poor pupil-teacher relationship”. Further, various risk domains with a significant medium overall effect (0.160 <  r  < 0.252) or a significant small effect ( r  < 0.160) were found (indicated in Table ​ Table2 2 with “ b ” and “ c ”, respectively).

Overall effect sizes of all risk domains for school absenteeism

# studies number of studies, # ES number of effect sizes, SE standard error, CI confidence interval, Sig significance, Mean z mean effect size (Fisher’s z ), % Var percentage of variance explained, Level 2 variance variance between effect sizes from the same study, Level 3 variance variance between studies, Mean r the correlation coefficient corresponding to the mean effect size z

Risk factors that could not be classified into one of the 44 created risk domains for school absenteeism are presented in italics

+ p  < 0.10; * p  < 0.05; ** p  < 0.01; *** p  < 0.001

a Significant large effect (according to the guidelines of Rice and Harris 2005 )

b Significant medium effect (according to the guidelines of Rice and Harris 2005 )

c Significant small effect (according to the guidelines of Rice and Harris 2005 )

For 15 domains, the estimated overall effect did not significantly deviate from zero implying that these domains cannot be regarded as risk domains given the present results. Of these 15 domains, three had as trend significant overall effect. Table ​ Table2 2 also shows the effects of 4 single factors (presented in italics ) that could not be classified in any of the created risk domains, due to their unique nature. The effect of the factors “history of grade retention”, “low attachment to parents”, and “no subculture affiliation” were significant and medium to small in size. The effect of “parental absenteeism in past” was not significant, implying that this variable was not identified as a risk factor for school absenteeism.

Overall Effects of Risk Domains for Dropout

Table ​ Table3 3 shows the overall effects of the 42 risk domains for school dropout. A significant effect in a positive direction was found for 23 risk domains, including 13 child-related domains, 7 family-related domains, 1 school-related domain, and 2 peer-related domains. Based on the criteria of Rice and Harris ( 2005 ), the magnitude of the significant overall effects ranged from small (i.e., r  = 0.062 for “ethnicity”) to large (i.e., r  = 0.365 for “history of grade retention”). Three child related risk domains with a large significant effect were found ( r  ≥ 0.299; indicated in Table ​ Table3 3 with “ a ”), including “history of grade retention”, “low IQ/learning difficulties”, and “low academic achievement”. Table ​ Table3 3 also lists several risk domains with a significant medium overall effect (0.192 <  r  < 0.299) or a significant small overall effect ( r  < 0.192) (indicated with “ b ” and “ c ”, respectively).

Overall effect sizes of all risk domains for school dropout

Risk factors that could not be classified into one of the 42 created risk domains for dropout are presented in italics

+ p   < 0.10; * p  < 0.05; ** p  < 0.01; *** p  < 0.001

The estimated overall effect did not significantly deviate from zero for 19 risk domains. This implies that these domains cannot be regarded as risk domains for dropout. Three of these 19 risk domains showed a trend significant effect. Table ​ Table3 3 also shows the overall effects of 6 single risk factors (presented in italics ). The factors “poor general well-being”, “adverse childhood experiences”, “age of mother (being younger)”, “large classes/schools” and “multicultural peer group” showed a significant medium to small overall effect size. The effect of the factor “sibling at school” was not significant, and could therefore not be identified as a risk factor for school dropout.

Table ​ Table4 4 presents the results of the three analyses that were conducted to assess bias in the estimated mean effect of each of the 43 risk domains for school absenteeism. There was no indication of bias in 13 estimated risk domain effects (i.e., 0 out of 3 methods indicated bias), some indication of bias in 22 risk domain effects (i.e., 1 out of 3 methods indicated bias), and moderate to strong indications of bias in 9 risk domain effects (i.e., 2 or 3 out of 3 methods indicated bias). These results show indications of bias in most of the estimated risk domains. For school dropout, no indication of bias was found in 14 estimated risk domain effects, some indication of bias in 20 risk domain effects, and moderate to strong indications of bias in 8 risk domain effects (see Table ​ Table5). 5 ). Again, an indication of bias was found in most risk domains. For brevity, the funnel plots that were produced in the trim-and-fill analyses are not presented here, but are available upon request from the first author.

Results of three methods for the assessment of bias in the estimated mean effects of the risk domains for school absenteeism

r mean effect size (Pearson”s correlation; see also Table ​ Table2), 2 ), Underestimation effect sizes were imputed to the right of the mean effect, implying that above average effect sizes were underrepresented and that the mean effect may be an underestimation of the true effect, Overestimation effect sizes were imputed to the left of the mean effect, implying that below average effect sizes were underrepresented and that the mean effect may be an overestimation of the true effect, NA not available, as only two or three effect sizes were classified in the risk domain

Dashes indicate that trimming and filling of effect sizes were not necessary according to the trim-and-fill algorithm

+ p <  0.10; * p  < 0.05; ** p  < 0.01; *** p  < 0.001

Results of three methods for the assessment of bias in the estimated mean effects of the risk domains for school dropout

r mean effect size (Pearson”s correlation; see also Table ​ Table3), 3 ), Underestimation effect sizes were imputed to the right of the mean effect, implying that above average effect sizes were underrepresented and that the mean effect may be an underestimation of the true effect, Overestimation effect sizes were imputed to the left of the mean effect, implying that below average effect sizes were underrepresented and that the mean effect may be an overestimation of the true effect, NA not available, because only two effect sizes were classified in the corresponding risk domain, or because all effect sizes classified in the risk domain originated from one study

The Moderating Effect of Gender

Table ​ Table2 2 shows the results of the likelihood-ratio tests that were performed to examine heterogeneity in effect sizes in the school absenteeism risk domains. In 37 risk domains, there was significant level-2 and/or level-3 variance. In the risk domains “psychiatric symptoms/disorders”, “low IQ/learning difficulties”, “large family size”, “sibling at school”, “distance to school (short)”, and “having many of friends”, there was no indication for heterogeneity in effect sizes. Therefore, no moderator analyses were performed in these domains. Further, and as mentioned in the Method section, the percentage of boys was only tested as a moderator when this variable was based on at least five studies. In the end, moderator analyses were performed for 20 risk domains for school absenteeism, and the results are presented in Table ​ Table6. 6 . A significant moderating effect was only found in the risk domain “drug abuse”, showing that the effect of this domain decreased as the percentage of boys in samples increased.

Results of testing gender as a potential moderator in various risk domains for school absenteeism

# Studies number of studies, # ES number of effect sizes, Mean z mean effect size (z), CI confidence interval, β 1 estimated regression coefficient, Level 2 variance residual variance between effect sizes from the same study, Level 3 variance residual variance between studies

a Omnibus test of all regression coefficients in the model

b p -Value of the omnibus test

Table ​ Table3 3 shows the results of the likelihood-ratio tests for the school dropout risk domains. Significant level-2 and/or level-3 variance was found in 32 risk domains. There was no indication for heterogeneity in effect sizes in the risk domains “delinquent behavior”, “not being religious”, “having a job”, “anxiety”, “large family size”, “parental alcohol use”, “history of child abuse victimization”, “negative school/class climate”, “often changed schools”, and “involvement with truant/deviant peers”. Also taking into account the lower bound that was set to five studies (see Method section), the percentage of boys was tested as a moderator in 15 risk domains for school dropout. The results are presented in Table ​ Table7, 7 , and reveal that only the overall effect of “having a negative school attitude” was moderated by gender. This finding implied that the effect of this risk domain for dropout decreased as the percentage of boys in samples increased.

Results of testing gender as a potential moderator in various risk domains for school dropout

+ p  < 0 .10; * p  < 0.05; ** p  < 0 .01; *** p  < 0.001

A great amount of literature has reported on potential risk factors for school absenteeism and/or school dropout, but a systematic review summarizing effects of risk factors for school absenteeism and risk factors for dropout was not yet available. Therefore, the aim of the present study was to estimate a mean effect of various risk domains (i.e., groups of more or less similar risk factors) for school absenteeism and various risk domains for school dropout. Both these constructs were examined in this meta-analytic review, as youths with excessive absenteeism are at high risk for permanent dropout from school (i.e., Kearney 2008a ) and therefore, the constructs may share various risk factors. However, it is also relevant to examine whether and how risk factors for school absenteeism differ from risk factors for school dropout. The second aim of this study was to examine whether the percentage of boys in samples moderates the overall strength of individual risk domains for school absenteeism or dropout.

Overall Effect of Risk Domains

The results revealed that multiple child-, family-, school- and peer-related risk factors contribute to the risk for both school absenteeism and school dropout. This is in line with the interdisciplinary model of school absenteeism formulated by Kearney ( 2008a ), in which several types of school absenteeism are influenced by various factors, including child, parent, family, peer, school, and community variables.

For school absenteeism, a significant overall effect was found for 28 out of 44 examined risk domains, ranging from r =  0.099 for having a low IQ or experiencing learning difficulties to r =  0.553 for having a negative school attitude. Large effects were found for 11 risk domains, including risks related to having a negative attitude towards school, substance abuse, externalizing and internalizing problem behavior of the child, and a low involvement of parents in school. For ease of interpretation, a number of “risk themes” were formulated that capture all significant risk domains (see also Assink et al. 2019 who applied this procedure in their review on risk factors for victimization of child sexual abuse). Given the current results, seven themes could be identified. First, the results indicate that moderate to large effects were found for multiple risk domains related to physical and mental problems of the child, which were: showing problematic internalizing behavior (other than being depressed and having anxieties; r =  0.307), having psychiatric symptoms or disorders ( r  = 0.303), being depressed ( r  = 0.237), having a poor physical health ( r  = 0.178), and suffering from anxieties ( r  = 0.115). Related to this theme, it was secondly found that risks referring to substance abuse had large effects, including smoking ( r =  0.336), drug abuse ( r  = 0.327), and alcohol abuse ( r =  0.311). Third, several antisocial or risky behaviors of the child were identified as risk factors, including showing anti-social behavior or having anti-social cognitions ( r =  0.428), a high sexual involvement ( r  = 0.229), showing risky behavior (such as risky behavior in traffic; r  = 0.226), and showing ineffective coping or having a risky personality profile ( r  = 0.158). Fourth, it was found that multiple risk domains relate to different sorts of problems at or with school , including having a negative school attitude ( r  = 0.503), a poor teacher-pupil relationship ( r  = 0.286), low levels of academic achievement ( r  = 0.232), a history of grade retention ( r  = 0.100), and a low IQ or learning difficulties ( r =  0.099). Related to this theme are different characteristics of the school that pose a risk for absenteeism, including a low quality of the school or education ( r =  0.229) and a negative school or class climate ( r  = 0.183). Sixth, parenting problems and difficulties are also important risk factors for school absenteeism, as significant effects were found of parents showing low levels of school involvement ( r  = 0.272), a low parental attachment ( r  = 0.220), parental mental or physical problems ( r =  0.186), low levels of parental support or acceptance ( r =  0.182), and low levels of parental control ( r  = 0.123). Finally, family (structure) problems could also be designated as significant risks, including a history of child abuse victimization in the family ( r  = 0.257), a non-nuclear family structure ( r  = 0.187), a low level of parental education ( r =  0.155), an ineffective family system ( r  = 0.154), and a low family SES ( r  = 0.134).

For school dropout, a significant overall effect in a positive direction was found for 23 out of 42 risk domains. Large effects were found for the risk factors having a history of grade retention ( r =  0.348), having a low IQ or experiencing learning difficulties ( r =  0.326) and showing low levels of academic achievement ( r =  0.316). For the dropout risk domains and the significant individual risk factors seven risk themes could be identified, with six themes being similar to those formulated for school absenteeism. First, problems at or with school were important risks for dropout. Medium to large effects were found for the risk domains having a history of grade retention ( r  = 0.348), having a low IQ or learning difficulties ( r  = 0.326), low levels of academic achievement ( r  = 0.316), and having a negative school attitude ( r =  0.210). The second risk theme consist of physical and mental problems of the child, such as: having psychiatric problems or disorders ( r  = 0.269), abusing drugs ( r =  0.247), poor general well-being ( r  = 0.210), having adverse childhood experiences ( r =  0.185), poor physical health ( r  = 0.157), and internalizing behavior problems (other than being depressed or having anxieties; r  = 0.140). Third, several anti-social behaviors were identified as risk factors for school dropout, including showing anti-social behavior or having anti-social cognitions ( r  = 0.236), engaging in delinquent behavior ( r  = 0.223), showing risky behaviors ( r =  0.109), and being involved with truant or deviant peers ( r  = 0.228). Fourth, parenting problems and difficulties were found to be important risk factors for school dropout, including low levels of parental support or acceptance ( r  = 0.176), low levels of parental involvement in school ( r  = 0.149), and low levels of parental control ( r =  0.134) Fifth, other family (structure) problems could be designated as significant risks, as significant effects were found for a low family SES ( r  = 0.222), a low educational level of parents ( r =  0.200), large families ( r =  0.194), and a non-nuclear family structure ( r =  0.178). Sixth, school dropout was related to characteristics of the school such as a negative climate in school or class ( r  = 0.147) and large schools or classes ( r =  0.145). Finally, the results showed that peer group characteristics or social status within a peer group had small significant effects on school dropout, including having many friends or being popular ( r  = 0.096) and being involved in a multicultural peer group ( r  = 0.088). This final risk theme is unique for school dropout. Naturally, the involvement with truant or deviant peers, which is was previously mentioned as part of the risk theme related to the anti-social behaviors of the child, can also be regarded as part of this final risk theme.

The abovementioned risk themes for school absenteeism and dropout are largely similar in nature, suggesting that both school absenteeism and dropout are related to similar risk factors. This was in line with what could be expected, because young people showing excessive absenteeism are at high risk for permanent school dropout. In his interdisciplinary model, Kearney ( 2008a ) suggests that several factors influence problematic school absenteeism, which could deteriorate over time from an acute, to a chronic, to a permanent state (dropout) of absenteeism. Moreover, since school drop-out is a more serious form of school absenteeism, it is possible that dropping out of school mainly results from an accumulation of multiple (different) risk factors, whereas the presence of a single (strong) risk factor may already lead to school absenteeism. This is also in line with the findings of Suh et al. ( 2007 ) indicating that as risk factors accumulate, students are more likely to drop out of school.

Moderating Effect of Gender

The variable percentage of boys in samples of primary studies was examined as a potential moderator of the overall strength of risk domains in which heterogeneity in effect sizes was identified. For school absenteeism, the effect of abusing drugs increased as the percentage of boys in samples decreased. This means that abusing drugs is a stronger risk factor for school absenteeism in girls than in boys. Previous research indicates that drug abuse rates are higher in men than in women (e.g., Becker and Hu 2008 ; Center for Behavioral Health Statistics and Quality 2017 ). Therefore, it is possible that drug abuse in boys is perceived as “more normal” or less deviant than in girls. This may imply that drug abuse contributes more to the risk of school absenteeism in girls than in boys.

For school dropout, it was found that only the effect of having a negative school attitude was moderated by the percentage of boys in primary study samples. The effect of this risk domain decreased as the percentage of boys increased, which means that having a negative school attitude is a stronger predictor of school dropout in girls than in boys. Prior research has revealed that boys have a more negative attitude towards school than girls (e.g., Harvey 1985 ; Logan and Johnston 2009 ). This negative attitude may stem from the fact that most school environments are centered around group and team work, whereas school environments in which autonomy is fostered (e.g., authority, aggression, and technical competence; Daniels et al. 2001 ) would better fit a masculine orientation to learning. As girls are generally less negative about school, it may be that girls with a negative school attitude may have to deal with other risk factors that are related to this negative attitude. Therefore, a negative school attitude might contribute more to the risk of school dropout in girls than in boys. It must be noted that most risk domains were not moderated by gender, indicating that the effect of most risk domains for school absenteeism and dropout seem similar for boys and girls.

Limitations

Several limitations of the present study should be mentioned. First, despite an extensive search procedure, it cannot be assured that the current sample of included studies is representative of all studies on (putative) risk factors for school absenteeism and dropout. A large amount of literature is available on the effect of risk factors for school absenteeism and dropout, and therefore it is possible that primary studies were missed. However, given the current extensive data set (a total of 69 studies and 1384 effect sizes), it may be assumed that the included studies were sufficiently representative of all primary studies available on risk factors for school absenteeism and dropout. Furthermore, the study inclusion was restricted to published studies and dissertations, there was a risk for overestimating effects of risk domains due to publication bias. The three tests for bias assessment indicated that bias may have been present in multiple estimated effects of risk domains. However, trim-and-fill analyses showed that an underestimation rather than an overestimation of risk domain effects was a problem (see Tables ​ Tables4 4 and ​ and5). 5 ). Therefore, bias in the analyzes data may not be due to specifically publication bias.

Second, the present study does not permit conclusions about causality between the presence of a risk factor and school absenteeism or dropout, because of the non-experimental nature of the included studies. In addition, in extracting effects of (putative) risk factors from primary studies, there was a focus on antecedents of school absenteeism and dropout (see also the inclusion criteria mentioned in the Method section), but as many included studies were retrospective in nature, it cannot be assured that all factors classified into the risk domains were true antecedents rather than outcomes. Further, it has been acknowledged that risk factors for school absenteeism and dropout are not present in isolation, but coexist and interact with other risk factors (e.g., Berends and Diest 2014; Ingul et al. 2012 ; Kearney 2008a , 2008b ). However, in the main focus of the present study was the mean effect of individual risk domains, and each risk factor was therefore classified into one of mutually exclusive risk domains. This allowed conducting a separate meta-analysis for each risk domain in order to estimate the mean effect of groups of (more or less) similar risk factors for school absenteeism and dropout. However, this did not allow us to examine what combinations of risk domains (or risk factors) may especially be predictive for school absenteeism and dropout. This may be a focus in future youth and adolescence research.

Finally, in the analytic strategy used this study, it was decided to only examine the variable percentage of boys in samples of primary studies as a potential moderator of risk domain effects. This decision was made as performing a large number of moderator analyses is not only impractical, but also statistically unwise, as insufficient data and capitalization on chance pose important problems. Furthermore, it was decided to only perform moderator analyses for variables that were based on at least five studies. Most coded variables did not meet this criterion, as some risk domains consisted of a small number of studies and effect sizes. As it was decided to only examine one potential moderator, the current study does not elaborate on the potential differences in overall effects of risk domains across different study designs or children with different background characteristics (e.g., age). Therefore, future youth and adolescence research should focus on examining effects of specific risk factors in different groups and under different circumstances.

Implications of the Study

The current study has a number of important implications. First, the current findings contribute to the fundamental knowledge of the etiology of school absenteeism and dropout, which in turn contributes to a better understanding of the problematic development of adolescents. Based on earlier research, it was already known that school absenteeism and dropout are caused by multiple child, parent, family, peer, and school factors. This study adds knowledge about which factors are most important in the etiology of both school absenteeism and dropout. This is important knowledge, for example for school professionals, that can be used in detecting risks of school absenteeism and dropout at an early stage, and in providing adequate prevention.

Furthermore, proper risk and needs assessment is essential in answering which children are at risk for school absenteeism or dropout and may therefore benefit from an (preventive) intervention. Risk and needs assessment may also indicate what factors should be targeted in an intervention so that the risk for school absenteeism or dropout could be reduced. So far, only measures have been developed to assess child factors among youth with specific types of school absenteeism, such as school refusal behavior (Kearney 2002 ; Kearney and Silverman 1993 ) and truancy (Kim and Barthelemy 2010 ). It was found that various child-, family-, school-, and peer-related risks are related to school absenteeism and dropout. Therefore, the results of this review show that the risk for school absenteeism and dropout can best be assessed from a multifactorial perspective in future risk- and need assessment instruments. This is in line with the suggestion of a multiaxial assessment of risk factors by Kearney ( 2008a ). Practitioners should focus on the assessment of factors related to the abovementioned risk themes, as it was found that these themes describe the risks that are predictive for school absenteeism and dropout. Furthermore, the risk domains with high overall effects on school absenteeism, including risks related to substance abuse and externalizing behavior, were most predictive and therefore deserve specific attention within risk- and need assessment instrument. Assessment instruments for school dropout should specifically focus on the child’s IQ, learning difficulties of the child, and a history of grade retention. As permanent dropout is often the consequence of excessive school absenteeism (Kearney 2008a ), it can be argued to assess both school absenteeism and dropout in a single instrument, while taking into account the differences in impact between school absenteeism risk factors and dropout risk factors. Furthermore, the findings of this review can be used to improve the validity of risk and needs assessment tools, as these findings indicate which risk factors are most strongly related to school absenteeism and dropout and should therefore be assessed by these tools. Assessing more relevant risk factors increases the validity of risk and needs assessment instruments.

As for the broad and multifactorial perspective that is needed in risk and needs assessment, (preventive) interventions should also be based on the notion that school absenteeism and dropout results from the presence of multiple child-, family-, school-, and peer-related factors. This means that all these factors should be taken into account in order to effectively reduce or prevent school absenteeism and dropout. Further, previous review studies indicate an insufficient effect of currently available intervention and preventions programs (Maynard et al. 2013 ; Wilson and Tanner-Smith 2013 ). This indicates a need for more effective interventions, for which the current findings may serve as a foundation.

School absenteeism and dropout are associated with many different life-course problems. To reduce the risk for these problems it is important to gain insight into risk factors for both school absenteeism and permanent school dropout. Until now, no quantitative overview of these risk factors and their effects was available. Therefore, this study was aimed at meta-analytically synthesizing the available evidence on risk factors for school absenteeism and dropout. The results of this study revealed that a substantial number of risks contribute to school absenteeism and dropout. For school absenteeism, significant and substantial effects were found for risks that refer to: physical and mental problems of the child (e.g., having psychiatric symptoms or disorders), substance abuse (e.g., drug abuse), antisocial or risky behavior (e.g., showing anti-social behavior or having anti-social cognitions), problems at or with school (e.g., having a negative school attitude), characteristics of the school (e.g., low quality of the school or education), parenting problems and difficulties (e.g., low parental school involvement), and family problems (e.g., an ineffective family system). As for school dropout, similar risks were identified next to risks related to peer group characteristics or social status in a peer group. The results imply that a multifactorial approach is needed in risk and needs assessment, and in interventions aimed at reducing or preventing school absenteeism and dropout. This review provides valuable insights for the development and improvement of both assessment and (preventive) intervention strategies.

Authors’ Contributions

J.G. participated in the design and coordination of the study, performed the statistical analysis, and drafted the manuscript; C.E.P. conceived of the study, participated in the coordination of the study and helped draft the manuscript; M.A. participated in the coordination of the study and helped draft the manuscript. All authors read and approved the final manuscript.

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Data Sharing and Declaration

The datasets generated and/or analyzed during the current study are not publicly available but are available from the corresponding author on reasonable request.

Biographies

is a PhD candidate at the Research Institute of Child Development and Education of the University of Amsterdam. Her research is mainly focused on risk assessment and prevention of child maltreatment and juvenile delinquency.

Claudia van der Put

is an assistant professor at the Research Institute of Child Development and Education of the University of Amsterdam. Her research interest is risk assessment and prevention of adverse developmental outcomes such as juvenile delinquency and problematic child-rearing situations.

is a postdoctoral researcher at the Research Institute of Child Development and Education of the University of Amsterdam. His research is mainly focused on developing and validating risk assessment instruments for child maltreatment and delinquency, determining risk factors for different negative developmental outcomes and examining the effectiveness of (preventive) psychosocial interventions.

Examples of risk factors classified in each risk domain

Note . The risk domains are in boldface; A = School absenteeism; D = School dropout

Compliance with Ethical Standards

The authors declare that they have no conflict of interest.

Because the research is a literature review, obtaining approval of an ethical committee was not necessary.

This research is a literature review, so these statements are not applicable.

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

  • *References marked with an asterisk were included in the meta-analysis
  • *Alexander KL, Entwisle DR, Horsey CS. From first grade forward: early foundations of high school dropout. Sociology of Education. 1997; 70 :87–107. [ Google Scholar ]
  • *Aloise-Young PA, Cruickshank C, Chavez EL. Cigarette smoking and perceived health in school dropouts: a comparison of Mexican American and Non-Hispanic white adolescents. Journal of Pediatric Psychology. 2002; 27 :497–507. [ PubMed ] [ Google Scholar ]
  • Andrews DA, Bonta J. The psychology of criminal conduct. Newark, NJ: Lexis/Nexis; 2010. [ Google Scholar ]
  • Andrews DA, Bonta J, Hoge RD. Classification for effective rehabilitation: rediscovering psychology. Criminal Justice and Behavior. 1990; 17 :19–52. [ Google Scholar ]
  • Andrews DA, Dowden C. A meta-analytic investigation into effective correctional intervention for female offenders. Forum on Corrections Research. 1999; 11 :18–21. [ Google Scholar ]
  • *Archambault I, Janosz M, Fallu J-S, Pagani LS. Student engagement and its relationship with early high school dropout. Journal of Adolescence. 2009; 32 :651–670. [ PubMed ] [ Google Scholar ]
  • Assink M, Van der Put CE, Meeuwsen MWCM, De Jong NM, Oort FJ, Stams GJJM, et al. Risk factors for child sexual abuse victimization: a meta-analytic review. Psychological Bulletin. 2019; 145 (5):459489.. [ PubMed ] [ Google Scholar ]
  • Assink M, Wibbelink CJM. Fitting three-level meta-analytic models in R: a step-by-step tutorial. The Quantitative Methods for Psychology. 2016; 12 :154–174. [ Google Scholar ]
  • *Attwood G, Croll P. Truancy in secondary school pupils: prevalence, trajectories and pupil perspectives. Research Papers in Education. 2006; 21 :467–484. [ Google Scholar ]
  • *Bask M, Salmela-Aro K. Burned out to drop out: exploring the relationship between school burnout and school dropout. European Journal of Psychology of Education. 2013; 28 :511–528. [ Google Scholar ]
  • *Battin-Pearson S, Newcomb MD, Abbott RD, Hill KG, Catalano RF, Hawkins JD. Predictors of early high school dropout: a test of five theories. Journal of Educational Psychology. 2002; 92 :568–582. [ Google Scholar ]
  • Becker JB, Hu M. Sex differences in drug abuse. Frontiers in Neuroendocrinology. 2008; 29 (1):36–47. [ PMC free article ] [ PubMed ] [ Google Scholar ]
  • Berends, I., & Van Diest, H. (2014). Schoolverzuim verklaard: Een overzicht van protectieve en risicofactoren [School absenteeism explained: an overview of protective and risk factors]. PI Research. https://www.piresearch.nl/files/2413/naar+een+verklaringsmodel+voor+schoolverzuim+juni+2014.pdf .
  • *Birkett M, Russell ST, Corliss HL. Sexual-orientation disparities in school: the mediational role of indicators of victimization in achievement and truancy because of feeling unsafe. American Journal of Public Health. 2014; 104 :1124–1128. [ PMC free article ] [ PubMed ] [ Google Scholar ]
  • *Blodgett C, Lanigan JD. The association between adverse childhood experience (ACE) and school success in elementary school children. School Psychology Quarterly. 2018; 33 (1):137–146. [ PubMed ] [ Google Scholar ]
  • *Blondal KS, Adalbjarnardottir S. Parenting in relation to school dropout through student engagement: a longitudinal study. Journal of Marriage and Family. 2014; 76 :778–795. [ Google Scholar ]
  • *Bobakova D, Geckova AM, Klein D, Van Dijk JP, Reijneveld SA. Fighting, truancy and low academic achievement in youth subcultures. YOUNG. 2015; 23 :357–372. [ Google Scholar ]
  • *Borgna C, Struffolino E. Pushed or pulled? Girls and boys facing early school leaving risk in Italy. Social Science Research. 2017; 61 :298–313. [ PubMed ] [ Google Scholar ]
  • *Bosker RJ, Hofman WHA. School effects on drop out: a multi-level logistic approach to assessing school-level correlates of drop out of ethnic minorities. Tijdschrift voor Onderwijsresearch. 1994; 19 :50–64. [ Google Scholar ]
  • *Breuner CC, Smith MS, Womack WM. Factors related to school absenteeism in adolescents with recurrent headache. Headache. 2004; 44 :217–222. [ PubMed ] [ Google Scholar ]
  • Bronfenbrenner U. The ecology of human development: Experiments by nature and design. Cambridge, MA: Harvard University Press; 1979. [ Google Scholar ]
  • Bronfenbrenner U. Ecology of the family as a context for human development. American Psychologist. 1986; 32 :513–531. [ Google Scholar ]
  • *Bryk AS, Thum YM. The effects of high school organization on dropping out: an exploratory investigation. American Educational Research Journal. 1989; 26 :353–383. [ Google Scholar ]
  • *Burton CM, Marshal MP, Chisolm DJ. School absenteeism and mental health among sexual minority youth and heterosexual youth. Journal of School Psychology. 2014; 52 :37–47. [ PMC free article ] [ PubMed ] [ Google Scholar ]
  • Center for Behavioral Health Statistics and Quality. (2017). Results from the 2016 National Survey on Drug Use and Health: Detailed Tables . Substance Abuse and Mental Health Services Administration. https://www.samhsa.gov/data/sites/default/files/NSDUH-DetTabs-2016/NSDUH-DetTabs-2016.pdf .
  • Cheung MWL. Modeling dependent effect sizes with three-level meta-analyses: a structural equation modeling approach. Psychological Methods. 2014; 19 :211–229. [ PubMed ] [ Google Scholar ]
  • Chou L-C, Ho C-Y, Chen C-Y, Chen WJ. Truancy and illicit drug use among adolescents surveyed via street outreach. Addictive Behaviors. 2006; 31 :149–154. [ PubMed ] [ Google Scholar ]
  • *Christle CA, Jolivette K, Nelson CM. School characteristics related to high school dropout rates. Remedial and Special Education. 2007; 28 :325–339. [ Google Scholar ]
  • *Claes E, Hooghe M, Reeskens T. Truancy as a contextual and school-related problem: a comparative multilevel analysis of country and school characteristics on civic knowledge among 14 year olds. Educational Studies. 2009; 35 :123–142. [ Google Scholar ]
  • Cooper H. Research synthesis and meta-analysis: A step-by-step approach. (5th edn). Thousand Oaks, CA: Sage; 2010. [ Google Scholar ]
  • *Corville-Smith J, Ryan BA, Adams GR, Dalicandro T. Distinguishing absentee students from regular attenders: the combined influence of personal, family, and school factors. Journal of Youth and Adolescence. 1998; 27 :629–648. [ Google Scholar ]
  • *Cratty D. Potential for significant reductions in dropout rates: analysis of an entire 3rd grade state cohort. Economics of Education Review. 2012; 31 :644–662. [ Google Scholar ]
  • Daniels H, Creese A, Hey V, Leonard D, Smith M. Gender and learning: equity, equality and pedagogy. Support for Learning. 2001; 16 (3):112–116. [ Google Scholar ]
  • De Baat, M., & Foolen, N. (2012). Oorzaken van schoolverzuim en voortijdig schoolverlaten [Causes of school absenteeism and dropping out of school]. Nederlands Jeugd Instituut. https://www.nji.nl/nl/Download-NJi/Oorzaken_schoolverzuimenvsv.pdf .
  • *Duncan DT, Hansen AR, Baidal JW, Lyn B, Hill A, Zhang J. Perceived not actual overweight is associated with excessive school absenteeism among U.S. adolescents. Obesity Research & Clinical Practice. 2017; 11 :398–405. [ PubMed ] [ Google Scholar ]
  • Dunkin MJ. Types of errors in synthesizing research in education. Review of Educational Research. 1996; 66 (2):87–97. [ Google Scholar ]
  • Durlak JA, Lipsey WM. A practitioner”s guide to meta-analysis. American Journal of Community Psychology. 1991; 19 :291–332. [ PubMed ] [ Google Scholar ]
  • *Dupéré, V., Dion, E., Leventhal, T., Archambault, I., Crosnoe, R., & Janosz, M. (2017). High school dropout in proximal context: the triggering role of stressful life events. Child Development . 10.1111/cdev.12792. [ PMC free article ] [ PubMed ]
  • Duval S, Tweedie R. A nonparametric ‘trim and fill’ method of accounting for publication bias in meta-analysis. Journal of the American Statistical Association. 2000; 95 (449):89–99. [ Google Scholar ]
  • Duval S, Tweedie R. Trim and fill: a simple funnel-plot-based method of testing and adjusting for publication bias in meta-analysis. Biometrics. 2000; 56 (2):455–463. [ PubMed ] [ Google Scholar ]
  • *Eaton DK, Brener N, Kann LK. Associations of health risk behaviors with school absenteeism. Does having permission for the absence make a difference? Journal of School Health. 2008; 78 :223–229. [ PubMed ] [ Google Scholar ]
  • *Echeverria SE, Velez-Valle E, Janevic T, Prystowsky A. The role of poverty status and obesity on school attendance in the United States. Journal of Adolescent Health. 2014; 55 :402–407. [ PubMed ] [ Google Scholar ]
  • *Egger HL, Costello JE, Angold A. School refusal and psychiatric disorders: a community study. Journal of the American Academy of Child & Adolescent Psychiatry. 2003; 42 :797–807. [ PubMed ] [ Google Scholar ]
  • *Ensminger ME, Lamkin RP, Jacobson N. School leaving: a longitudinal perspective including neighborhood effects. Child Development. 1996; 67 :2400–2416. [ PubMed ] [ Google Scholar ]
  • *Ensminger ME, Slusarcick AL. Paths to high school graduation or dropout: a longitudinal study of a first-grade cohort. Sociology of Education. 1992; 65 :95–113. [ Google Scholar ]
  • *Epstein JL, Sheldon SB. Present and accounted for: improving student attendance through family and community involvement. Journal of Educational Research. 2002; 95 :308–318. [ Google Scholar ]
  • *Fall AM, Roberts G. High school dropouts: interactions between social context, self-perceptions, school engagement, and student dropout. Journal of Adolescence. 2012; 35 :787–798. [ PMC free article ] [ PubMed ] [ Google Scholar ]
  • Ferguson GA. Statistical analysis in psychology & education. New York, NY: McGraw-Hill; 1966. [ Google Scholar ]
  • *Fernandez-Suarez, A., Herrero, J., Perez, B., Juarros-Basterretxea, J., & Rodríguez-Díaz, F. J. (2016). Risk factors for school dropout in a sample of juvenile offenders. Frontiers in Psychology , 7 . 10.3389/fpsyg.2016.01993. [ PMC free article ] [ PubMed ]
  • *Flisher AJ, Chalton DO. High-school dropouts in a working-class South African community: selected characteristics and risk-taking behavior. Journal of Adolescence. 1995; 18 :105–121. [ Google Scholar ]
  • *Fortin L, Lessard A, Marcotte D. Comparison by gender of students with behavior problems who dropped out of school. Procedia Social and Behavioral Sciences. 2010; 2 :5530–5538. [ Google Scholar ]
  • Fremont WP. School refusal in children and adolescents. American Family Physician. 2003; 68 :1555–1560. [ PubMed ] [ Google Scholar ]
  • *Gastic B. School truancy and the disciplinary problems of bullying victims. Educational Review. 2008; 60 :391–404. [ Google Scholar ]
  • *Gleason P, Dynarski M. Do we know whom to serve? Issues in using risk factors to identify dropouts. Journal of Education for Students Placed at Risk. 2002; 7 :25–41. [ Google Scholar ]
  • *Gleason, P., & Dynarski, M. (1994). Falling behind: characteristics of students in federally funded prevention programs . https://www.mathematica-mpr.com/download-media?MediaItemId=%7B37670860-FF8D-46D3-AE90-7655E3038D0D%7D .
  • *Hagborg JM, Berglund K, Fahlke C. Evidence for a relationship between child maltreatment and absenteeism among high-school students in Sweden. Child Abuse & Neglect. 2018; 75 :41–49. [ PubMed ] [ Google Scholar ]
  • Hammond, C., Linton, D., Smink, J., & Drew, J. (2007). Dropout risk factors and exemplary programs: a technical report . National Dropout Prevention Center/Network. https://files.eric.ed.gov/fulltext/ED497057.pdf .
  • *Hancock KJ, Mitrou F, Taylor CL, Zubrick SR. The diverse risk profiles of persistently absent primary students: implications for attendance policies in Australia. Journal of Education for Students Placed at Risk. 2018; 23 (1-2):53–69. [ Google Scholar ]
  • Harvey TJ. Gender differences in attitudes to science and school for first year secondary school children in a variety of teaching groups. Educational Review. 1985; 37 (3):281–288. [ Google Scholar ]
  • *Henry KL. Who’s skipping school: characteristics of truants in 8th and 10th grade. Journal of School Health. 2007; 77 :29–35. [ PubMed ] [ Google Scholar ]
  • *Hickman GP, Bartholomew M, Mathwig J, Heinrick RS. Differential developmental pathways of high school dropouts and graduates. Journal of Educational Research. 2008; 102 :3–14. [ Google Scholar ]
  • Houben M, Van den Noortgate W, Kuppens P. The relation between short-term emotion dynamics and psychological well-being: a meta-analysis. Psychological Bulletin. 2015; 141 :901–930. [ PubMed ] [ Google Scholar ]
  • *Hughes, J. N., West, S. G., Kim, H., & Bauer, S. S. (2018). Effect of early grade retention on school completion: a prospective study. Journal of Educational Psychology , (7), 974-991. 10.1037/edu0000243. [ PMC free article ] [ PubMed ]
  • *Hunt MK, Hopko DR. Predicting high school truancy among students in the Appalachian south. Journal of Primary Prevention. 2009; 30 :549–567. [ PubMed ] [ Google Scholar ]
  • *Hysing M, Haugland S, Stormark KM, Boe T, Sivertsen B. Sleep and school attendance in adolescence: results from a large population-based study. Scandinavian Journal of Public Health. 2015; 43 :2–9. [ PubMed ] [ Google Scholar ]
  • *Ingul JM, Klõckner CA, Silverman WK, Nordahl HM. Adolescent school absenteeism: modelling social and individual risk factors. Child and Adolescent Mental Health. 2012; 17 :93–100. [ PubMed ] [ Google Scholar ]
  • *Ingul, J. M., & Nordahl, H. M. (2013). Anxiety as a risk factor for school absenteeism: what differentiates anxious school attenders from non-attenders? Annals of General Psychiatry , 12 . 10.1186/1744-859X-12-25. [ PMC free article ] [ PubMed ]
  • Jaafar NRN, Iryani MDT, Salwina WIW, Nazri ARF, Kamal NA, Prakash RJ, et al. Externalizing and internalizing syndromes in relation to school truancy among adolescents in high-risk urban schools. Asia-Pacific Psychiatry. 2013; 5 :27–34. [ PubMed ] [ Google Scholar ]
  • *Janosz M, LeBlanc M, Boulerice B, Tremblay RE. Disentangling the weight of school dropout predictors: a test on two longitudinal samples. Journal of Youth and Adolescence. 1997; 26 :733–762. [ Google Scholar ]
  • *Jenkins PH. School delinquency and school commitment. Sociology of Education. 1995; 68 (3):221–239. [ Google Scholar ]
  • *Jimerson S, Egeland BL, Sroufe A, Carlson B. A prospective longitudinal study of high school dropouts examining multiple predictors across development. Journal of School Psychology. 2000; 38 :525–549. [ Google Scholar ]
  • Kearney CA. Identifying the function of school refusal behavior: a revision of the School Refusal Assessment Scale. Journal of Psychopathology and Behavioral Assessment. 2002; 24 :235–245. [ Google Scholar ]
  • Kearney CA. An interdisciplinary model of school absenteeism in youth to inform professional practice and public policy. Educational Psychology Review. 2008; 20 :257–282. [ Google Scholar ]
  • Kearney CA. School absenteeism and school refusal behavior in youth: a contemporary review. Clinical Psychology Review. 2008; 28 :451–471. [ PubMed ] [ Google Scholar ]
  • Kearney CA, Silverman WK. Measuring the function of school refusal behavior: The School Refusal Assessment Scale. Journal of Clinical Child Psychology. 1993; 22 :85–96. [ Google Scholar ]
  • *Keppens G, Spruyt B. Dé spijbelaar bestaat niet: een empirisch onderzoek naar types van occasionele spijbelaars in Vlaanderen [The typical truant does not exist: an empirical research on different categories of occasional truants in Flanders] Mens en Maatschappij. 2015; 90 :143–169. [ Google Scholar ]
  • Kim H, Barthelemy JJ. A tool for assessing truancy risk among school children: predictive and construct validity of the Risk Indicator Survey. Journal of Social Service Research. 2010; 37 :50–60. [ Google Scholar ]
  • King NJ, Bernstein GA. School refusal in children and adolescents: a review of the past 10 years. Journal of the American Academy of Child and Adolescent Psychiatry. 2001; 40 :197–205. [ PubMed ] [ Google Scholar ]
  • Knapp G, Hartung J. Improved tests for a random effects meta-regression with a single covariate. Statistics in Medicine. 2003; 22 :2693–2710. [ PubMed ] [ Google Scholar ]
  • Kogan SM, Luo Z, Murry VM, Brody GH. Risk and protective factors for substance use among African American high school dropouts. Psychology of Addictive Behaviors. 2005; 19 :382–391. [ PubMed ] [ Google Scholar ]
  • Kraemer HC, Kazdin AE, Offord DR, Kessler RC, Jensen PS, Kupfer DJ. Coming to terms with the terms of risk. Archives of General Psychiatry. 1997; 54 :337–343. [ PubMed ] [ Google Scholar ]
  • Lipsey MW, Wilson DB. Practical meta-analysis. Thousand Oaks, CA: Sage; 2001. [ Google Scholar ]
  • *Lloyd DN. Prediction of school failure from third-grade data. Educational and Psychological Measurement. 1978; 38 :1193–1200. [ Google Scholar ]
  • Logan S, Johnston R. Gender differences in reading ability and attitudes: examining where these differences lie. Journal of Research in Reading. 2009; 32 (2):199–214. [ Google Scholar ]
  • *Lounsbury JW, Steel RP, Loveland JM, Gibson LW. An investigation of personality traits in relation to adolescent school absenteeism. Journal of Youth and Adolescence. 2004; 33 :457–466. [ Google Scholar ]
  • Maynard BR, McCrea KT, Pigott TD, Kelly MS. Indicated truancy interventions for chronic truant students: a Campbell systematic review. Research on Social Work Practice. 2013; 23 :5–21. [ Google Scholar ]
  • *Mounteney J, Haugland S, Skutle A. Truancy, alcohol use and alcohol-related problems in secondary school pupils in Norway. Health Education Research. 2010; 25 :945–954. [ PubMed ] [ Google Scholar ]
  • *Mullvain, P. (2017). Examining the relationship between bullying, attendance, and achievement in schools (Doctoral dissertation). http://scholarworks.waldenu.edu/cgi/viewcontent.cgi?article=3717&context=dissertations .
  • National Center for Education Statistics. 2015 and 2017 Mathematics and Reading Assessments. Washington, DC: US Department of Education; 2018. [ Google Scholar ]
  • *Peguero AA, Ovink SM, Li YL. Social bonding to school and educational inequality: race/ethnicity, dropping out, and the significance of place. Sociological Perspectives. 2016; 59 :317–344. [ Google Scholar ]
  • *Quiroga CV, Janosz M, Lyons JS, Morin AJS. Grade retention and seventh-grade depression symptoms in the course of school dropout among high-risk adolescents. Psychology. 2012; 3 :749–755. [ Google Scholar ]
  • R Core Team. (2015). R: A language and environment for statistical computing . R Foundation for Statistical Computing. http://www.R-project.org/ .
  • *Ramberg J, Låftman SB, Fransson E, Modin B. School effectiveness and truancy: a multilevel study of upper secondary schools in Stockholm. International Journal of Adolescence and Youth. 2019; 24 (2):185–198. [ Google Scholar ]
  • *Ramirez M, Wu Y, Kataoka S, Wong M, Yang J, Peek-Asa C, Stein B. Youth violence across multiple dimensions: a study of violence, absenteeism, and suspensions among middle school children. The Journal of Pediatrics. 2012; 161 :542–546. [ PubMed ] [ Google Scholar ]
  • *Reid K. Alienation and persistent school absenteeism. Research in Education. 1981; 26 :31–40. [ Google Scholar ]
  • *Reid K. Some social, psychological and educational aspects related to persistent school absenteeism. Research in Education. 1984; 31 :63–82. [ Google Scholar ]
  • Rice ME, Harris GT. Comparing effect sizes in follow-up studies: ROC Area, Cohen”s d, and r. Law and Human Behavior. 2005; 29 :615–620. [ PubMed ] [ Google Scholar ]
  • Rosenthal, R. (1994). Parametric measures of effect size. In H. Cooper & L. V. Hedges (Eds.), The Handbook of Research Synthesis (pp. 239). New York, NY: Sage.
  • *Rousseau-Salvador C, Amouroux R, Annequin D, Salvador A, Tourniaire B, Rusinek S. Anxiety, depression and school absenteeism in youth with chronic or episodic headache. Pain Research and Management. 2014; 19 :235–240. [ PMC free article ] [ PubMed ] [ Google Scholar ]
  • *Rumberger RW. Dropping out of middle school: a multilevel analysis of students and schools. American Educational Research Journal. 1995; 32 :583–625. [ Google Scholar ]
  • *Rumberger RW, Ghatak R, Poulos G, Ritter PL, Dornbusch SM. Family influence on dropout behavior in one California high school. Sociology of Education. 1990; 63 :283–299. [ Google Scholar ]
  • *Sälzer C, Trautwein U, Lüdtke O, Stamm M. Predicting adolescent truancy: the importance of distinguishing between different aspects of instructional quality. Learning and Instruction. 2012; 22 :311–319. [ Google Scholar ]
  • *Sapharas NK, Estell DB, Doran KA, Waldron M. Effects of parental divorce or a father’s death on high school completion. Psychology in the Schools. 2016; 53 :861–874. [ Google Scholar ]
  • *Schwartz LA, Radcliffe J, Barakat LP. Associates of school absenteeism in adolescents with sickle cell disease. Pediatric Blood and Cancer. 2009; 52 :92–96. [ PMC free article ] [ PubMed ] [ Google Scholar ]
  • *Sommer B. What’s different about truants? A comparison study of eighth-graders. Journal of Youth and Adolescence. 1985; 14 :411–422. [ PubMed ] [ Google Scholar ]
  • *Steinhausen, H-C., Müller, N., & Metzke, C. W. (2008). Frequency, stability and differentiation of self-reported school fear and truancy in a community sample. Child and Adolescent Psychiatry and Mental Health , 2 . 10.1186/1753-2000-2-17. [ PMC free article ] [ PubMed ]
  • Suh S, Suh. J, Houston I. Predictors of categorical at-risk high school dropouts. Journal of Counseling & Development. 2007; 85 :196–203. [ Google Scholar ]
  • *Sznitman SR, Reisel L, Khurana A. Socioeconomic background and high school completion: mediation by health and moderation by national context. Journal of Adolescence. 2017; 56 :118–126. [ PubMed ] [ Google Scholar ]
  • Tabachnik BG, Fidell LS. Using multivariate statistics. (6th edn). Boston, MA: Allyn and Bacon; 2013. [ Google Scholar ]
  • *Taylor, R. G. (2009). School facilities in the nation’s capital: an analysis of student achievement, attendance, and truancy (Doctoral dissertation). http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.620.4220&rep=rep1&type=pdf .
  • Teasley ML. Absenteeism and truancy: risk, protection, and best practice implications for school social workers. Children & Schools. 2004; 26 :117–128. [ Google Scholar ]
  • *Teuscher S, Makarova E. Students’ school engagement and their truant behavior: Do relationships with classmates and teachers matter? Journal of Education and Learning. 2018; 7 (6):124–137. [ Google Scholar ]
  • *Torney-Purta, J., Lehmann, R., Oswald, H., & Schulz, W. (2001). Citizenship and education in twenty-eight countries: civic knowledge and engagement at age fourteen. https://files.eric.ed.gov/fulltext/ED452116.pdf .
  • Tramontina S, Martins S, Michalowski MB, Ketzer CR, Eizirik M, Biederman J, et al. School dropout and conduct disorder in Brazilian elementary school students. Canadian Journal of Psychiatry. 2001; 46 :941–947. [ PubMed ] [ Google Scholar ]
  • *Trampush JW, Miller CJ, Newcorn JH, Halperin JM. The impact of childhood ADHD on dropping out of high school in urban adolescents/young adults. Journal of Attention Disorders. 2009; 13 :127–136. [ PMC free article ] [ PubMed ] [ Google Scholar ]
  • Tyrrell M. School phobia. Journal of School Nursing. 2005; 21 :147–151. [ PubMed ] [ Google Scholar ]
  • Van den Noortgate W, López-López JA, Marin-Martinez F, Sánchez-Meca J. Three-level meta-analysis of dependent effect sizes. Behavior Research Methods. 2013; 45 :576–594. [ PubMed ] [ Google Scholar ]
  • Van den Noortgate W, López-López JA, Marin-Martinez F, Sánchez-Meca J. Meta-analysis of multiple outcomes: a multilevel approach. Behavior Research Methods. 2014; 47 :1274–1294. [ PubMed ] [ Google Scholar ]
  • Van der Woude LB, Van der Stouwe T, Stams GJJM. Differences between risk factors for truancy and delinquency in Dutch adolescents. Children and Youth Services Review. 2017; 73 :9–14. [ Google Scholar ]
  • *Vaughn MG, Maynard B, Salas-Wright C, Perron BE, Abdon A. Prevalence and correlates of truancy in the US: results from a national sample. Journal of Adolescence. 2013; 36 :767–776. [ PMC free article ] [ PubMed ] [ Google Scholar ]
  • *Veenstra R, Lindenberg S, Tinga F, Ormel J. Truancy in late elementary and early secondary education: the influence of social bonds and self-control—the TRAILS study. International Journal of Behavioral Development. 2010; 34 :302–310. [ Google Scholar ]
  • Viechtbauer W. Conducting meta-analyses in R with the metafor package. Journal of Statistical Software. 2010; 36 :1–48. [ Google Scholar ]
  • *Vitaro F, Larocque D, Janosz M, Tremblay RE. Negative social experiences and dropping out of school. Educational Psychology. 2001; 21 :401–415. [ Google Scholar ]
  • Wilson SJ, Tanner-Smith EE. Dropout prevention and intervention programs for improving school completion among school-aged children and youth: a systematic review. Journal of the Society for Social Work and Research. 2013; 4 :357–372. [ Google Scholar ]
  • *Zhang D, Katsiyannis A, Barrett DE, Willson V. Truancy offenders in the juvenile justice system. Remedial and Special Education. 2007; 28 :244–256. [ Google Scholar ]

Advertisement

Advertisement

A methodology to design, develop, and evaluate machine learning models for predicting dropout in school systems: the case of Chile

  • Published: 23 January 2023
  • Volume 28 , pages 10103–10149, ( 2023 )

Cite this article

research proposal school dropout

  • Patricio Rodríguez   ORCID: orcid.org/0000-0002-7361-5132 1 ,
  • Alexis Villanueva 2 ,
  • Lioubov Dombrovskaia 3 &
  • Juan Pablo Valenzuela 1  

3748 Accesses

6 Citations

Explore all metrics

School dropout is a structural problem which permanently penalizes students and society in areas such as low qualification jobs, higher poverty levels and lower life expectancy, lower pensions, and higher economic burden for governments. Given these high consequences and the surge of the problem due to COVID-19 pandemic, in this paper we propose a methodology to design, develop, and evaluate a machine learning model for predicting dropout in school systems. In this methodology, we introduce necessary steps to develop a robust model to estimate the individual risk of each student to drop out of school. As advancement from previous research, this proposal focuses on analyzing individual trajectories of students, incorporating the student situation at school, family, among other levels, changes, and accumulation of events to predict dropout. Following the methodology, we create a model for the Chilean case based on data available mostly through administrative data from the educational system, and according to known factors associated with school dropout. Our results are better than those from previous research with a relevant sample size, with a predictive capability 20% higher for the actual dropout cases. Also, in contrast to previous work, the including non-individual dimensions results in a substantive contribution to the prediction of leaving school. We also illustrate applications of the model for Chilean case to support public policy decision making such as profiling schools for qualitative studies of pedagogic practices, profiling students’ dropout trajectories and simulating scenarios.

Similar content being viewed by others

research proposal school dropout

Design, Development, and Evaluation of a Predictive Model for Regular School Dropout in the Chilean Educational System

research proposal school dropout

On predicting school dropouts in Egypt: A machine learning approach

research proposal school dropout

Dropout Early Warning System (DEWS) in Malaysia’s Primary and Secondary Education: A Conceptual Paper

Avoid common mistakes on your manuscript.

1 Introduction

School failure has a central place in educational systems due to its enduring effects on students. This happens when the school system fails to ensure that students reach certain levels of schooling, experiencing grade repetition and temporary or definitive dropout from school (OECD, 2020 ). This results in adults with low qualifications or students who do not complete their schooling at the secondary level.

School failure has moved from a vision that sees school dropout as a problem associated with the students’ —attributing responsibility to them— to one that understands it as an expression of a systemic problem where school system and society are also responsible (OECD, 2010 ).

1.1 Impact of school failure and its impact on students

As evidence states, school dropout permanently penalizes students and the whole society in aspects such as:

Low qualification jobs, lower skills to face the labor world and lower productivity (Gil et al., 2019 ; Lee-St John et al., 2018 ; Sahin et al., 2016 ) and higher unemployment (Lee-St John et al., 2018 ; Sahin et al., 2016 ).

Lower income (Lee-St John et al., 2018 ), higher poverty level and lower life expectancy (Sahin et al., 2016 ),

Lower pensions (Dussaillant, 2017 ) and higher economic burden to the State for social protection concepts (Höfter, 2006 ; Levin et al., 2012 ).

Higher crime rates (Lee-St John et al., 2018 ; Sahin et al., 2016 ), lower social cohesion and citizen participation (Sahin et al., 2016 ).

Lower economic growth and − in social terms − lower tax payments (Gil et al., 2019 ; Lee-St John et al., 2018 ).

1.2 Dropping out in the world

School failure is a structural problem in most societies. In OECD, the average percentage of adults between 25 and 64 years old whose maximum level of education is lower secondary OECD is 27% and Chile is slightly above with 35%. However, many other countries such as Colombia, Mexico, and Spain report higher dropout rates (OECD, 2020 ), as shown in Fig.  1 .

figure 1

Lower secondary as higher level attained for 25–64-year-old adults, showing differences by gender. Own elaboration based on OECD ( 2020 ) data

Additionally, due to COVID-19 pandemic, about 24 million learners, from pre-primary to university level, are at risk of not returning to school following the education disruption (UNESCO, 2020 ). For this, societies must address proactively all the drivers of educational exclusion to strengthen the resilience of education systems in the face of this crisis (UNESCO, 2020 ).

1.3 Risk factors for school dropout

The causes of dropout are associated with the students, their family, the school, educational system, and elements of the context or social environment where they are (Boniolo & Najmias, 2018 ; Weybright et al., 2017 ). Since this is a gradual and cumulative process, some indicators warn of this disengagement risk even from early in the school trajectory (Boniolo & Najmias, 2018 ). We can classify these risk factors at:

Individual level: such as school repetition and overage (Boniolo & Najmias, 2018 ; Hirakawa & Taniguchi, 2021 ), school attendance (Hirakawa & Taniguchi, 2021 ; Sahin et al., 2016 ), academic performance (Hirakawa & Taniguchi, 2021 ) and specific learning needs (Gil et al., 2019 ). Socioemotional factors are also included, like attitudes towards learning (Hirakawa & Taniguchi, 2021 ; Sahin et al., 2016 ; Zaff et al., 2017 ), non-academic problematic behaviors (Weybright et al., 2017 ) and school mobility (Sahin et al., 2016 ). As well as sociodemographic factors, such as gender, ethnicity, and nationality (Hirakawa & Taniguchi, 2021 ; Lee-St John et al., 2018 ).

Family level: such as socioeconomic status and parental involvement (Adelman et al., 2018 ; Boniolo & Najmias, 2018 ; Lee-St John et al., 2018 ; Sahin et al., 2016 ).

School level: its characteristics, socioeconomic and sociocultural composition (Hirakawa & Taniguchi, 2021 ), its resources (Dussaillant, 2017 ; Ecker-Lyster & Niileksela, 2016 ), the relationship between students and teachers (Gil et al., 2019 ) and participation in school activities (Gil et al., 2019 ).

Extra-school level considers community factors, e.g., the geographic location of residences, families and the condition of their housing, access to playgrounds, green areas, or “urbanity” (Zaff et al., 2017 ), having a network of high-achieving and aspirational peers (Hirakawa & Taniguchi, 2021 ). And it also includes contextual factors, understood as potential “pull factors” that incentivize early job attachment (Kattan & Székely, 2017 ).

1.4 Predictive models for school dropout based on machine learning

Machine Learning is a discipline that employs algorithms to automate tasks like classification and regression. Algorithms learn from known datasets (training sample) to estimate the true value of a target variable using predictor variables. To evaluate the performance of these results, they are contrasted with the real values in out-of- sample data (test or validation sample) (Sorensen, 2019 ).

A meta-analysis on academic literature and case studies on machine learning applications to predict dropout between 2013 and 2017 found that algorithms such as neural networks or decision trees are mainly used for the dropout prediction as a binary classification exercise on the dropout/non-dropout dichotomy (Mduma et al., 2019 ).

For predicting school dropout, researchers chose algorithms from the family of decision trees such as CART (Jena & Dehuri, 2020 ), and decision trees ensembles (Bentéjac et al., 2021 ). Sorensen ( 2019 ) elaborated a decision tree model to estimate, considering records in students’ last year of primary education, dropout in the secondary level predicting 63.3% of actual dropout cases using only academic and individual factors.

Similar dropout prediction models have been used to develop Early Warning Systems (EWS). These systems allow decision makers to identify in time students at risk of dropping out, to react to this notification and, eventually, to help potential dropouts to continue with their learning processes at different levels (Lee & Chung, 2019 ).

The main difficulty in large-scale dropout prediction is related to the severe imbalance of the phenomenon (Lee & Chung, 2019 ). Therefore, it is necessary to apply corrections, choosing an adequate model performance evaluation metrics and selecting a machine learning algorithm whose flexibility allows overfitting reduction (Lee & Chung, 2019 ; Sansone, 2019 ).

1.5 Purpose and structure of this article

Given the high consequences and impact of school dropout and the surge of the problem due the school closure during COVID pandemic (Khan & Ahmed, 2021 ; Pereira de Souza et al., 2020 ), our objective is to develop a methodology to design, develop, and evaluate a machine learning model for predicting dropout in school system. The aim of this methodology is supporting and guiding models’ development by practitioners and policy makers, − specially from Latin American and African countries (UNESCO, 2020 ) where the student dropout is higher than other regions − to implement national or subnational Early Warning Systems (EWS) to identify student with higher risk of abandoning their studies.

This methodology produces necessary steps to develop a robust model to estimate the individual risk of each student to drop out of school, generating applications to support public policy decision making. As advancement from previous research, this proposal focuses on analyzing individual trajectories of students, incorporating the student situation at multiple levels (school, family, among others), and changes and accumulation of events to predict dropout. In this way, we shift the computational from the machine learning model to the trajectories’ calculation, what is, a one-time development comparing to multiple trainings of models. Since machine learning model are less transparent (Sorensen, 2019 ) in this paper we provide a reliable option to explain results and how they depend on the context.

We develop a model for the Chilean educational system to illustrate a practical case, which is relevant for three reasons. First, reduction of school dropout has been a policy for the last decade; second, data quality permits sophisticated analysis for machine learning approach; and finally, Chile is a medium income country, therefore this experience could be useful for similar countries or others with less development level.

This paper is structured as follows. In Section 2 , we present the methodology defining the dimensions of robustness and describing every step to develop a predictive model. In Section 3 , we address the model development for the Chilean education system presenting the main results. In Section 4 , we present public policy applications, to end discussing the implications of this methodology in Section 5 .

2 A methodology for predicting school dropout using machine learning

The aim of this methodology is to produce a robust model to estimate the individual student risk for dropping out of school, to answer the research question stated in Section 1.5 . A robust model is one which fulfill the following criteria (Studer et al., 2021 ):

Has good general performance in the chosen metrics, allowing practical use in the context of application.

Is stable: the performance doesn’t depend on assumptions, imputed data, creation of training and test samples and has good general avoiding under and overfitting.

Is computational effective: has reasonable computational times for training and prediction, depending on the context of application.

Is easy to maintain, requiring the minimum variables to predict results, allowing to obtain data, and creating every case straightforward for training and predicting purposes.

Its explanations are consistent with the dominion of the model. The variables’ importance and its variance explanations are consistent with literature about the topic.

The methodology comprises the following steps, as they are shown in Fig.  2 .

figure 2

Steps of the proposed methodology for design, development and evaluating a machine learning model for predicting dropout. Each step answers specific questions about model robustness and produces an outcome for the next phase

In the following subsections, we will describe every step.

2.1 Step 1: Creating student trajectories

We define as the objective of the model to predict the first time where students leave their school (regular dropout) to avoid consequences stated in Section 1.1 . Therefore, every case should be codified to train the algorithm. Contemplating the risk factors identified previously (Section 1.3 ), we propose to codify each student’s history in the school system through a continuous range of years in a single vector of data. This is because each trajectory comprises the student situation at multiple levels (school, family, among others), and changes and accumulation of events are relevant to predict dropout (Kattan & Székely, 2017 ). Thus, we can identify clearly both dropping out and protective factors throughout the educational cycle after 12 years.

Every case should be labeled as dropout (1, positive class) or not (0, negative class) to use binary classification. Since we can verify in the data if they in a given year ( \(i\) ) are not enrolled in any school in the next year ( \(i+1\) ). In that case, we label that student as a dropout in year \(i\) .

Several datasets should be considered to incorporate dropout factors at individual, family, school, and extra school levels. This will facilitate an explanation of the model in step 6. How many variables associated with these factors will depend on the availability and reliability of data in the school system, being the most important challenge to face in the first place.

To reduce errors, data should be carefully cleaned. If there are several sources of data, we should perform several consistency analyses to ensure reliability of data: e.g., consistency of date of birth, sex, enrollment in schools in each year through the period analyzed. If there is some data that cannot be found, and we need to impute it (e.g., results of surveys of income and education of parents to create a socioeconomic status), we need to analyze the impact of chosen imputation methods on results in step 4.

2.2 Step 2: Creating training and test samples and choose performance metrics

For this kind of problems were there are a temporal prediction, we will not use a traditional sample construction which divide all the cases in a proportion such as 80% for training and 20% for testing. In this instance, we have dropouts until year \(t\) , and we need to predict if a given students will leave school at year \(t+1\) . Therefore, the sample and testing samples will follow the same logic (Sorensen, 2019 ).

Dropping out is a phenomenon naturally imbalanced since significantly fewer students abandon school than graduate. Thus, specific solutions for training unbalanced data should be used (Mduma et al., 2019 ). Two options to deal with the imbalance are undersampling the majority class (non-dropouts) or oversampling the minority class (dropouts).

Since each case codifies a student trajectory, we propose to create the training and test samples as follows. To include in the training sample cases to compare the variables from trajectories that lead to regular dropouts with those that don’t, we create a set of counterfactual trajectories for each student who drops out. For each student who drops out in a grade-year ( \(i,m\) ), we generate trajectories belonging to students of the same cohort Footnote 1 who don’t drop out or do at a later grade-year ( \(i+j,m+n\) ) with \(j\ge 0,n\ge 1\) . Hence, all these trajectories were abbreviated multiple times based on their counterfactual similarity (evaluated by year and grade reached) of a dropout case (see Fig.  3 ).

figure 3

Example of two counterfactual trajectories for student A, who begins primary education in 2004 and dropouts in 6 th grade in 2009. Since students B and C (which also began education in 2004) graduate, or dropout in a higher grade and year, their counterfactual trajectories are calculated to the same grade and year of student A

Subsequently, we will create a training sample until a year \(t\) considering all the dropouts at year \(i\) ( \(i\le t\) ) plus a random undersampling of the total counterfactual trajectories, creating a sample where dropout cases have higher proportion than the natural prevalence of the phenomenon, e.g., 30%, 40% or 50%. Additionally, some stratification criteria to subsample contrafactual trajectories can be used also based on some variables such as grade, sex, socioeconomic status, or schools’ categories to ensure their representativity. The impact of these stratification procedures on results should be also tested on step 4.

For the test sample, all the trajectories which should be on school in a year \(t+1\) are included to measure the performance of the model prediction. Test sample remains unbalanced.

Imbalanced models should be also evaluated with appropriate performance metrics. We consider the following performance metrics for binary classification:

Recall : is the class hit rate with respect to the total number of real cases belonging to that class. The false negative rate is 1 − recall. Minimizing false negatives, it ensures students could potentially drop out are detected.

Precision : is the class hit rate with respect to the total number of predictions for a class. The false positive rate is 1 − precision.

Sensitivity : Recall of the positive class in a binary classifier.

Specificity : Recall of the negative class in a binary classifier.

F1 score: Harmonic mean between the precision and the recall of a class, in this case, the positive one.

To balance results of true positive and negative rates, we use the geometric mean between the sensitivity and specificity (GM Score) to measure performance in the test sample (Márquez-Vera et al., 2016 ). Also, we consider the recall and precision of both classes, and the F1 score of the positive class (Mduma et al., 2019 ).

2.3 Step 3: Selecting predictive algorithms and train model

In the third step, we choose an algorithm to train the model. There are several algorithms to create the model such as Decision Trees and its ensembles, SVM machines, neural networks between others (Şara et al., 2015 ). However, international experiences in the application of machine learning for the prediction of school dropout strongly suggest the use of decision tree ensemble algorithms since:

They are better suited to deal with both continuous and categorical variables (Jena & Dehuri, 2020 ).

They have shown a robust performance in exercises of a similar nature (Lee & Chung, 2019 ; Sansone, 2019 ; Sorensen, 2019 ).

Ensemble decision trees use strategies to avoid overfitting (Bentéjac et al., 2021 ).

In this regard, there are a set of decision trees with gradient boosting ensembles such as eXtreme Gradient Boosting ( XGBoost ) (Chen & Guestrin, 2016 ) , Light Gradient Boosting Machine ( LightGBM ) (Ke et al., 2017 ) and Categorical Boosting ( CatBoost ) (Prokhorenkova et al., 2019 ) .

2.4 Step 4: Results and sensitivity analysis

The performance obtained with the test sample, should be carefully analyzed to discard under or overfitting. Overfitting occurs where performance is very good with the training sample but bad with test sample, and underfitting when performance is bad in both samples (Dos Santos et al., 2009 ).

In previous steps, several imputing methods and assumptions were made, and the quality of the obtained trajectories may vary. Other decisions include how stratified random undersampling methods were used to create the training sample. Impact of these assumptions on the stability of the results should be tested.

Also, model could have better results in some contexts, for example in some levels or categories of schools (e.g., public vs private, urban vs rural). These contexts can determine the limitations of the model or where could be used with more confidence.

2.5 Step 5: Model improvement

As outcome from step 4, we obtained an initial model. If such a model has good results in terms of performance the question which arises is: can we refine our model making it easy to maintain and with better computer performance in the training and predicting tasks? Producing a model easy to main means reducing the quantity of variables involved, finding a subset which enables us to make predictions at the same level of performance metric. Less variables will reduce both the effort to create the trajectories for the training and simplify obtaining data for prediction.

There are at least three algorithms to discard and determine relevance of each variable in estimating results:

Naïve Recursive Feature Elimination (RFE): eliminates variables recursively until the minimum number that maximizes the performance of the model (given an objective function) is obtained (Misra & Yadav, 2020 ). In this case, we propose to use the GM score as an objective function (see Section  2.2 ).

Boruta: evaluates the importance of each variable with respect to a permuted version of it to determine its relevance (Kursa et al., 2010 ).

Shap RFE: is a modified version of RFE that identifies more robustly the importance of each variable using SHAP (see Section  2.6 ) (Lundberg et al., 2019 ; Sharma et al., 2020 ).

Alternatively, each machine learning algorithm has some parameters (hyperparameters) which can affect both performance results and computation time to train the model. There are optimization hyperparameters algorithms based on brute force, Bayesian statistics, genetic algorithms, among others. In the present work, we consider two of them:

Tree Parzen Estimator (TPE): is a semi-random optimization algorithm improving performance by analyzing the history of parameters already used, seeking the optimization of a loss function based on Parzen Estimators (Bergstra et al., 2011 ).

Population Based Training (PBT): is an evolutionary mechanism where generations of hyperparameter configurations are created. Then, PBT evaluates their performance and selects the best ones, creating a new generation of configurations with changes with respect to previous one, repeating the process until algorithm stops after a given number of iterations, or no improvement appears (Jaderberg et al., 2017 ).

A last option for model improvement is using a corrective model to decrease the false positive rate after variable reduction and hyperparameters optimization. Reducing false positive rate decreases students identified wrongly as possible leavers. For this, a second model is trained with true positives and false positives results of previous years. Thus, to correct the original model to predict dropping out on \(t+1\) , the corrective model is trained until year \(t-1\) . These two models are applied successively and if both agree that students are positive cases, then the overall results are positive as shown in Fig.  4 .

figure 4

Improvement of the optimized model created in step 4, using a second corrective model

2.6 Step 6: Explaining the model

For the public accountability, no discrimination and transparency criteria in decision making where automatic systems are involved should be fulfilled (Buenadicha et al., 2019 ). Algorithmic discrimination refers where discrimination occurring in real world is reproduced in data environments, e.g., by gender or ethnic. Algorithm transparency refers to data they collect, how they manage it, how they analyze it, with whom they share it, what decisions are made based on it and based on what factors.

Therefore, after the best model is obtained, an explanatory model to understand how the model makes its predictions should be created. We propose using SHAP ( SHapley Additive exPlanations ) (Lundberg et al., 2019 ), because this method allows us to estimate the contribution of each variable to individual predictions in a robust, consistent, and locally accurate way (Lundberg et al., 2018 ). It uses an optimized procedure for tree-based algorithms allowing interpreting and debugging the resultant model (Sharma et al., 2020 ; Yoshida, 2020 ).

Thus, the output of this explanatory model is the probability decomposed into the specific contribution of each variable. Hence, all variable contributions for a given student sums his/her probability of dropping out. Using this method, students and schools can be profiled based on the contribution of each variable in the final probability of dropout (Section  4 ).

3 Using proposed methodology to Chilean education system

For better understanding of the model development, we first present an overview of the Chilean education system (Section  3.1 ). From Section  3.2 to 3.7 , we develop a model for the Chilean case.

3.1 Overview of Chilean education system

In Chile, compulsory education lasts 12 years. Grades 1 st to 8 th are for primary education and grades 9 th to 12 th are for secondary education, with three cycles: first cycle (1 st – 4 th grade), second cycle (5 th – 8 th grades) and third cycle (9 th – 12 th grades).

In the early 1980s, Chile implemented a school choice system, introducing a per- student subsidy mechanism (voucher scheme). The per-student subsidy is the same for public and private schools meant to cover the school’s operating costs. Students can attend the school of their choice without administrative boundaries restrictions. This policy was supposed to stimulate competition between schools to attract and retain students, leading to improved efficiency and higher quality educational services (Ladd & Fiske, 2020 ).

There are mainly three school categories of schools Footnote 2 : a) public schools , funded by the per-student subsidy paid by the state and run by each of 345 municipalities, b) private-voucher schools , funded by the per-student subsidy paid by the state and operated by the private sector, and c) private fee- paying schools , financed solely by fees paid by parents, and run by the private sector (Ladd & Fiske, 2020 ).

An ongoing system-wide reform in public school education calls for de-municipalization of the public-school sector. This creates 70 new Local Education Services (LES) between 2018 and 2025, consolidating administration of schools formerly under mayoral control (Anderson et al., 2021 ). These 70 LES respond to a new agency responsibility for public schools: The Directorate of Public Education.

Figure  5 shows global dropout incidence rate. The highest one appears when students transit from primary to secondary education.

figure 5

Global dropout per level in 2019, showing differences by gender. Grades are shown according to International Standard Classification of Education (UNESCO, 2012 ). Source: Own Elaboration based on Chilean Ministry of Education open data

Students can enroll in adult education, which accepts over 15-year-olds to primary or over 17-years-olds to secondary education. Since also they can enroll in more than one grade per year, this is a de facto alternative to avoid regular dropout. National evidence shows enrollment in this modality increases at higher grades. In 2019, 57,130 students left regular education to adult modality. In contrast, only 36,230 students dropped out the same year.

3.2 Step 1: Creating student trajectories

With the available datasets, we created student trajectories as an analyzable artifact. This consists of three procedures to produce a single observation summarizing a time ordered sequence of each student's transit from the first grade of regular primary education to the last reported period. These procedures are: 1) Dataset collection, 2) Determining student sequences; and 3) summarizing sequences into trajectories.

3.2.1 Datasets collection

In previous works on measuring and predicting school dropout, the data were collected from surveys and using administrative sources to obtain longitudinal data. For example, Sorensen ( 2019 ) and Lee and Chung ( 2019 ) used data from administrative and secondary sources to identify and quantify variables associated with students' situations.

We chose to use administrative data, obtained from secondary sources collected, organized, and published by Ministry of Education (MINEDUC) since 2004 in its open data platform, and to a lesser extent, from data related to the Education Quality Measurement System (SIMCE Footnote 3 ) census tests as well as the parents and students’ surveys made available by the Agency for Quality in Education (AQE) for research purposes (Table 1 ).

In these datasets, all students are identified anonymously by masking their National Identification Number (Masked ID or MID). Thus, individual data can be cross-referenced, and we can trace the trajectories of every student.

Using administrative data for assembling student trajectories is a great opportunity to identify trends and patterns that lead to dropout. Still, some limitations need to be considered. Mainly the exclusion of certain factors identified as relevant in the literature, but difficult to measure or non-existent in administrative data, like, e.g., contextual factors or non-academic problem behaviors.

When we consolidated all this data, we found several inconsistencies through the years, such as implausible birth years, data inconsistencies for the same student, gender discrepancies, academic statuses reported without enrollments and vice versa, students skipping grades, students graduated on a non-final grade and mismatches on the grade reported in SIMCE/PDSI datasets.

To address these problems, these situations were operationalized to subsequently assess the consistency of the reported history for each student (see Table 2 ).

3.2.2 Determining students’ sequences

We generate sequence tables composed of time-ordered series, for each student, where the student's situation is described with respect to their trajectory: enrollment status, dropout incidence, among others. For each student in the sequence table, we calculate new variables relevant to the model such as changes of grade, changes of school between or within the same year and grade repetitions. The procedure is as follows:

We standardized the available information on enrollment and academic status. These operate as articulating axes of the sequences to allow traceability. Then, we completed them by assigning data from other sources.

We assigned each student to a cohort from the first available period (2004).

After cohort assignment, the base sequences are created with enrollment and academic status of the students, matching enrollment-academic status pairs available for each year and MID. When there is no either enrollment or academic status, fictitious enrollment and academic status data are created by duplicating the available case and filling in the unavailable columns with missing values.

When there is more than one enrollment or academic status, we define the following criteria to identify the unique enrollment- academic status pair to represent the period within the sequence (Table 2 ).

After performing the above procedure, it was possible to trace the sequences of each student in the cohorts from the year 2004, obtaining sequence tables. However, it was only possible to create trajectories for students entering the first year of primary education in 2004 and, therefore, the number of students per year whose trajectory is feasible increases each successive year and stabilizes after 12 years, when the students of the 2004 cohort reached their last grade.

Even after 12 years, it is not possible to create the trajectory of all students because it is not possible to identify their cohort of origin, which occurs, for example, in the case of foreigners who do not start school in the national system, however, a traceability rate of 96.4% is achieved.

3.2.3 Summarizing sequences into student trajectories

Given the amount of data available, we opted for traditional supervised machine learning methods to generate the predictive model over larger scale alternatives traditionally used in forecasting exercises, such as models based on neural networks. This is because such models work with training samples larger than those available, and they have lower interpretability.

Then, we reduced the sequence tables of each student to a single observation describing their passage through school education. To this end, we generate a student trajectory, adding variables created from grouping the sequences table, summarizing the student's final situation, their most frequent values in some variables (for instance, number of public schools attended) and other elements related to risk factors identified in the literature. The socioeconomic status is included in the family level risk factor, and it was calculated as the mean of the standardized declared household income and the maximum standardized parental schooling. In this case, a multilevel imputation was performed to deal with the high number of missing cases. Additionally, we included other sequence descriptors, allowing us to capture relevant milestones of the trajectory summarized, such as the last year or grade reached.

Since the original raw data contains inconsistencies resulting from the data collection procedures, we create a score to evaluate the quality of the trajectories and analyze the consequences of considering s with lower consistency. We define 15 inconsistency indicators in 3 levels: from the data reported on enrollment and performance (10 indicators, level 1), from datasets provided by MINEDUC (2 indicators, level 2) and from data reported by other sources (3 indicators, level 3) (complete criteria are available in Table 9 in Appendix  A ). The consistency score was normalized with mean 0 and standard deviation 1, the distribution per cohort is shown in Fig.  6 .

figure 6

Inconsistencies per cohort. Every color line represents an inconsistency level that increases by severity. The consistency of the student trajectories by cohort has gradually improved in the last 10 years making levels two or three infrequent. Values are in log 10 scale

Each trajectory operationalizes regular dropout, which is where a student enrolled in some grade for children and youth on year \(t\) is either enrolled in adult education or out of the school system on year \(t+1\) . We also include consistency descriptors to control and evaluate the quality with which the trajectories are calculated with the procedure described in Section  2.3 . Thus, 111 variables were considered and grouped according to its type (Table 3 ), including SIMCE and IDPS ones. Complete variable descriptions are available in Appendix B .

Using administrative data limits the availability of contextual or family variables compared to more readily available individual and school data. At the end, we generated 3,847,469 student trajectories.

Using these trajectories, a first visual exploratory analysis allows us to recognize differences of performance on dropout of the different schools on LES territories, by school dependency and total school enrollment (Fig.  7 ). As Fig.  7 shows, there are territories with performance worse than regression predicted and should be the focus of public policies.

figure 7

Regression models built based on students’ trajectories by total enrollment in school categories. Several public schools on LES territories underperform. The private-voucher schools in the same territory have better results with greater enrollment

3.3 Step 2: Training and test samples and performance metrics

As we explained in Section  2.2 , we create a training and test sample with trajectories until 2018, and 2019 respectively.

The training sample was created with data until 2018. This sample is imbalanced since contains 3,847,469 trajectories of which 345,874 (8.9%) lead to dropout, with an imbalance ratio of 10.12. To deal with this problem, as it is proposed in Section  2.2 , we create contrafactual trajectories for students who drop out. We opted for a stratified subsampling using four variables: gender, category of the last school, last year and registered grade. This reduced the negative class from 26,793,262 counterfactuals to 345,874, amount equal to the number of dropouts. Therefore, the training sample has 691.748 cases, with an imbalance ratio of 1.

The test sample uses all student trajectories that reached 2019 and were (or not) dropouts in 2020, totaling 2,802,156 trajectories of which 47,632 (1.7%) lead to dropout with an imbalance ratio of 57.82. There is no intersection between training and test samples since their variables were constructed until different years. We will report recall and precision for each class, looking for better performance in GM and F1 scores. Given the sample sizes, it is unnecessary to use cross validation.

3.4 Step 3: Selecting predictive algorithms

We produced machine learning models using a basic decision tree algorithm as the simplest model and then we also tried 3 decision tree ensemble algorithms with gradient boosting: XGBoost , LightGBM and CatBoost (see Section  2.3 for justification). We trained and tested them using the samples created in the previous section.

3.5 Step 4: Results and Sensitivity analysis

The results of the four algorithms on the test sample considering 103 variables without missing data (excluding SIMCE and PSDI scores) are shown in Table 4 .

The performance of the tree algorithms with gradient boosting is superior to the classic CART decision tree. LightGBM is slightly superior to CatBoost in GM score. This indicates that while the CatBoost model achieves better performance in terms of recall for class 1 (which means fewer false negatives), it also has a higher false positive rate. Hereafter, the LightGBM model will be referenced as the base model.

3.5.1 Stability of performance on trajectories consistency, training sample creation and SIMCE and PDSI scores

Several models were also trained considering the internal consistency score of each trajectory. We concluded that it is necessary to consider all cases since less consistent trajectories also indicate a higher prevalence of dropout and discarding them does affect the final performance of the model.

The stability of the performance was evaluated for 100 random different samples of contrafactual trajectories. The greater variation was just 0.014 for the F1-score, as can be seen in Fig.  8 .

figure 8

Distribution and range of model performance results in multiple training samples

Variables related to performance on SIMCE tests and PDSI scores were also considered. but their contribution to the performance of the model was very low compared to the cost of obtaining these datasets and the high amount of missing data.

3.5.2 Performance in different grades, schools’ categories, and sizes

Since the base model is stable in trajectory consistency, random choice of counterfactuals and, SIMCE and PDSI variables did not introduce significant performance improvements, we finally address the question of how the base model performs in different contexts, defined by combinations of grades and school categories, thinking in its practical use (Table 5 ).

Performance tends to improve at higher grades (where the natural prevalence of dropping out in Chile is higher, as Fig.  5 shows) and in public schools (as shown in Fig.  7 ), with the best performance in secondary for public schools, and the worst relative performance in primary education in the private fee-paying sector. Despite that, these results are better than those from previous research with a relevant sample size (Lee & Chung, 2019 ), and a predictive capability 20% higher for the actual dropout cases, also considering the advantage of addressing the problem of classroom imbalance.

If we analyze the classification error of this model based on school size, public and private voucher schools follow the same patterns. Figure  9 shows the results for public schools, where error is minimal for false negative rates of any school size and false positive rates decreasing for large schools from 500 students.

figure 9

False negative rates ( left ) and false positive rates ( right ) by public school size for secondary and 7 th to 12 th grade

3.6 Step 5: Model improvement

To determine the relevance of each variable in estimating each student’s dropout probability and because the naïve version of RFE has problems in dealing with noise from irrelevant variables, we used two feature selection methods (see Section  2.5 ): 1) first using Boruta and then applying naïve RFE and 2) applying ShapRFE . Both approaches proved more effective than using naïve RFE, which discarded only 64 of the 103 original variables in contrast to the proposed methods which discarded 83 and 87 respectively. In Table 6 , we compare the performance of the three models with 103, 30 and 26 variables.

Both methods allowed to create simpler and more efficient models maintaining performance. Table 7 shows the contribution of the variables selected by the two previous methods by its type (as in Table 3 ), considering two values: 1) the aggregated contribution, which is the sum of the importance of each variable in the set, and 2) the average contribution, which is the aggregated contribution divided by the number of variables per type. The contributions of only the 26 relevant variables per type are depicted in Appendix B .

Individual level factors made the greatest contribution, consistent with literature. School factors and trajectories’ descriptors are also relevant in both approaches. Therefore, in contrast to previous work, the inclusion of non-individual dimensions results in a substantive contribution to the prediction of school dropout.

Since our final model is just trained in 16 s on a desktop computer (Table 6 ) and it takes less than 1 s to predict 2.8 million cases, we considered this a reasonable performance, and we did not optimize the hyperparameters of the LightGBM algorithm.

Finally, we generated a corrective model using the procedure described in Section  2.5 . For that, we took the 26 variables and identified true and false positives until 2018. False positives were codified as 0 and true positives as 1. The results of the correction are shown in Table 8 .

In all grades and school categories, recall scores for class 0, precision scores for class 1 and all the F1 scores improved to the minimal detriment of recall scores for class 1. In terms of absolute quantities, for secondary education in public schools, false positives diminished from 6.1% to 4.9% and for private voucher schools came from 6.16% to 5.15%. In the case of 7 th – 12 th grades, the reduction was from 6.35% to 5.25% in public schools, and from 6.1% to 5.19% in private voucher schools.

3.7 Step 6: Explaining the model

The SHAP method decomposes each individual probability prediction into the specific contribution of each variable. Thus, all variable contributions for a given student sums their probability of dropping out. SHAP values were computed from the initial model without the false positive correction. Figure  10 shows the contribution of each variable of the final model for two cases, one where the model predicts a high probability (0.99) and other a lower one (0.01).

figure 10

Individual variable contribution for two cases using SHAP values. SHAP values per variable for a student with low dropout probability (0.01) are shown in green , while SHAP values for a student with high dropout probability (0.99) are shown in red

Figure  11 shows the individual contribution of each variable selected for the final model with 26 variables for all 2019 cases.

figure 11

Beeswarm plot for final model. Every point shows the impact of each variable in a dropout prediction per student in the 2019 test sample. The colors denote the value of the variable in its own scale (high values in red, low ones in blue). Absolute mean contributions to predictions are ordered from left (higher) to right (lower). Variable codes are in Appendix B

For example, LAST_GRADE_APPRVD is a binary variable indicating if a student passed (1) or not (0) their last year at school. Figure  11 shows us two things: 1) LAST_GRADE_APPRVD is the most important variable in predicting dropping out and, 2) in all the cases lower LAST_GRADE_APPRVD values (0, shown in blue) have a positive contribution while higher values (1, shown in red) have negative contribution to dropping out probability. This analysis can be repeated for each school, allowing to identify the most important variables for dropout at local level. For example, the school on Fig. 12a  has a 0.5% dropout rate while the school on Fig.  12b ) has a 26.6% dropout rate. Variable importance ranks are different between schools and contribute in different ways.

figure 12

Beeswarm profiles of two schools with different dropout rates. The colors denote the value of the variable in its own scale (high values in red, low ones in blue). The school on the left (a) has a dropout rate of 0.5% while the school on the right (b) has 26.6%. Absolute mean contributions to predictions are ordered from above (higher) to below (lower). Variable codes are in Appendix B

Further implications for public policy will be discussed on Section  4.1 .

4 Public policy applications

The straightforward application of this model is developing an EWS. But as it was stated in Section  1.5 , we can also envision other applications of these models for decision and public policy making. These are: profiling schools for qualitative studies of pedagogic practices, profiling students’ dropout trajectories and simulating scenarios.

4.1 Profiling schools for qualitative studies of pedagogic practices

As it was stated in Section  2.6 , model explanations allow to guide further qualitative research about pedagogic practices. Results of the explained model can guide qualitative studies in schools. For example, the variable CL_STUDENT is a binary one indicating if a student is Chilean (1) or not (0) (see Appendix B ). As general results of the test sample show (Fig.  11 ), being foreign student increases your chance of dropping out. If we analyze a school with a low rate of dropout (Fig.  12a ), it is indifferent if a student is Chilean or not since the contribution of the variable to the dropout probability is negative. However, in the school of Fig.  12b , CL_STUDENT is the variable with most importance and being foreign has a positive contribution. Therefore, pedagogic practices with foreign students can be investigated further in both schools, and the question which arises is: what are the pedagogic practices that can be replicated (school a) or avoided (school b) in similar contexts?

Additionally, any significant difference in the quantities of dropouts expected at school or LES level could be indicative of changes in local policies for school retention with better or worse results.

4.2 Profiling students’ dropout trajectories

In second place, since SHAP values for every variable are continuous, we used clustering algorithms to identify typologies of trajectories leading to dropout. We used the 39,844 true positives’ SHAP values calculated for the year 2019 in a clustering model.

SHAP values were rescaled to adjust them to a range between -1 and 1, preserving the directionality of the predictions, but normalizing the different impact level of every variable. Since the excessive dimensionality of the data (26 variables), we used UMAP (McInnes et al., 2020 ) to reduce the information to only two. From this, 20 clusters were found using DBSCAN (Ester et al., 1996 ). The detailed characterization of clusters based on the original domain of each variable can be found in Table 10 in Appendix.

There are 3 main categories of trajectories: 1) where students completed and approved their last level (23.5%); 2) where students completed their last level but did not approve (30.4%), and 3) where students did not complete last level (46.1%). As Fig.  13 shows, within these 3 categories there are also subcategories based on just 5 variables: student is Chilean (CL_STUDNT), PPS beneficiary (PPS_BENFNC), Overage (OVRAGE), Last grade on school (LAST_GRADE) and Number of abandonments in the last cycle (NUM_ABN_LAST_CYCL). Category 1 has 5 clusters; Category 2 has 6 clusters and Category 3 has 9 clusters.

figure 13

Typology of trajectories based on the clustering model. Every square indicates a division by the variable indicated. Bifurcation to left is to lesser values and to the right to greater values

The 67% of the students’ trajectories are concentrated just in seven clusters (Table 9 ): one of category 1 (cluster 6), two of category 2 (clusters 0 and 5) and four of category 3 (clusters 1,4,10 and 13). In these clusters predominates the school categories, grade and sex expected according to incidence of the phenomenon (see Figs. 5 and 7 ). They have the following characteristics (from greater to lesser trajectories) according to variables in Table 12 in Appendix  C :

Cluster 4 (13.0%): Last grade not completed, mainly students in 9 th grade, almost only Chileans, with averages: SES of 0.22, attendance in last cycle of 82.13%, z score of − 0.99, repetition of 2.1, changes of schools of 2.47 and last school effectiveness of 41.96%.

Cluster 5 (11.7%): Last grade completed but nor approved, mainly students in 9 th and 10 th grade, almost only Chileans, with averages: SES of 0.24, attendance in last cycle of 75.7%, z score of − 1.64, repetition of 2.87, changes of schools of 2.51, last school effectiveness of 43.42%.

Cluster 1 (11.6%): Last grade not completed, mainly students in 10 th and 11 th grade, almost only Chileans, with averages SES of 0.31, attendance in last cycle of 85.37%, z score of − 0.63, repetition of 0.55, changes of schools of 2.09, last school effectiveness of 45.57%.

Cluster 0 (9.4%): Last grade completed but nor approved, mainly students in 9 th grade, almost only Chileans, with averages: SES of 0.29, attendance in last cycle of 74.52%, z score of − 1.66, repetition of 1.56, change of schools of 1.91, last school effectiveness of 46.26%.

Cluster 6 (7.3%): Completed and approved, mainly students in 7 th and 8 th grade, almost only Chileans, with averages: SES of 0.25, attendance in last cycle of 85.35%, z score of − 0.88, repetition of 2.20, change of schools of 2.42, last school effectiveness of 43.51%.

Cluster 10 (7.1%): Last grade not completed, mainly students in 4 th grade, only Chileans, with averages: SES of 0.47, attendance in last cycle of 88.53, z score of − 0.08, repetition of 0.16, change of schools of 0.59, last school effectiveness of 50%.

Cluster 13 (7.0%): Last grade not completed, only foreigners in 1 st and 2 nd grade, almost only Chileans, with averages: SES of 0.33, attendance in last cycle of 92.56%, z score of + 0.05, repetition of 0.03, change of schools of 0.08, last school effectiveness of 45.26%.

In all these clusters, the last schools were predominant public except in clusters 0 (48.6% voucher vs 42.7% public schools), 1 (44.5% voucher vs 44.2% public schools) and 10 (45.6% voucher vs 32.7% public schools). Apparently, the clustering model grouped in Cluster 13, all trajectories of foreign students which changed their identity number from a provisional to the official one. This causes an abnormal incidence of dropout in first and second grades since these trajectories were truncated by an administrative anomaly.

As Sansone ( 2019 ) verified, the heterogeneity of students at risk of dropping out through this kind of unsupervised learning, allowing to identify subpopulations among students and, thus, to design programs appropriate to each group, understanding both their peculiarities and key factors associated with their situation, so that policymakers could benefit from exploiting this to customize the treatment of each cluster of students.

4.3 Simulating scenarios: External shocks

In third place, predictive models can be used to evaluate impact in dropping out of external shocks, such as an economic recession, natural catastrophe, or a pandemic.

In this case, we present a simulation of the effect of a pandemic such as the COVID-19 in the increase of dropping out following the methodology described on Fig.  14 . The shock is created by applying scenario assumptions which alter the input data (scenario data), and the results of the model are compared in a base case (unchanged data) with the scenario data. Since we know the prediction error of the model (see Tables 5 and 8 ), we can correct final quantities to avoid overestimation.

figure 14

Procedure for simulating a scenario using the predictive model. In stage 1, the prediction is created as business usual (base). In step 2, the original dataset is altered according to the scenario assumptions, creating a modified dataset which is used for prediction (scenario). In step 3, since the error of prediction is known, the results are corrected using that generating a difference on dropout

For illustrative purposes, we analyzed the effect of diminishing attendance in the marginal increase of dropout. If we assume that all variables behave the same as 2019 where students attended in person and we just correct replace individual LAST_CYCL_AVG_ATTNDNC variable by a fixed factor, we obtain results shown in Fig.  15 .

figure 15

Results of a simulating of decreasing attendance by a given factor in additional dropouts. Note that factor a zero factor conduces to repetition, but not necessarily to dropout

On first semester of 2022, monthly attendance data from Mineduc shows that it is approximately 9% lower (equivalent to a factor of 0.91) in average compared to 2018 and 2019, for either public or private voucher schools. Therefore, without any intervention and this tendency remains and does not worsen, the simulation estimates 10,501 additional dropouts at end of the year 2022.

5 Discussion

In this paper, we proposed a methodology to design, develop, and evaluate a predictive model for regular school dropout using: 1) individual student trajectories as individual cases; 2) procedures for creating training and test samples, and choosing performance metrics considering class imbalance; 3) machine learning algorithms for this kind of problems; 4) sensitivity analyses to test dependency of results on previous assumptions, and determine contexts where the model works better; 5) methods to reduce variables improving maintenance and reducing false positives, and; 6) explanatory techniques to calculate the individual contribution of each variable to dropout probability.

Following the methodology, we develop a model for the Chilean case (Section  3 ) based on data available mostly through administrative data from the educational system, and according to known factors associated with school dropout. Our results are better than those from previous research with a relevant sample size (Lee & Chung, 2019 ), with a predictive capability 20% higher for the actual dropout cases. Also, in contrast to previous work, the inclusion of non-individual dimensions results in a substantive contribution to the prediction of leaving school. Contrary to Sorensen ( 2019 ), who found better results using SVM, Gradient based boosting decision trees worked best for us. Therefore, the importance of trying different algorithms in step 3.

Long-term policies can be devised to manage risk factors, such as academic lag, for reducing that prevalence in future cohorts of students. At school level, the model can identify students with higher dropout risk requiring support and protection strategies to ensure positive school trajectories. For example, those who have recently repeated, have high levels of absenteeism, have accumulated more than one repetition and are over-aged. In Chile, this is exacerbated when the student is male, migrant or has started his education overseas. Results show that these efforts will have greater impact in public schools, with lower socioeconomic levels from secondary education. As can be seen, these are all indicators that are easy to construct at school level. Also, in the case of Chile, these analyses will be useful for the Directorate of Public Education to understand the challenges of the territories that will become part of new public education soon.

The major contributions of this study are:

As Sorensen ( 2019 ) states, machine learning is less transparent and technological demanding. However, techniques like SHAP proposed in this paper provides a reliable option to explain results and how they depend on the context. Cloud computing infrastructure also reduces significantly computational cost, but, in our case, it was not necessary. This is because the burden of computational cost is shifted from the machine learning model to the trajectories’ calculation, what is, a one-time development comparing to multiple trainings of models.

The public policy applications envisioned in Section  4 , to inform public policies such as profiling schools for qualitative studies of pedagogic practices, profiling students’ dropout trajectories and simulating the impact of events such as pandemics or natural disasters. Simulations estimate the decreasing/increasing of dropout, providing information for calculating the return of investment of public policies on school retention.

Some limitations of this study are that the administrative nature of the available data limits the possibility of transforming the prediction into concrete action and, at the same time, gives a constrained vision of the school trajectory. In addition, since certain data are hard to obtain, it is difficult to assess their potential contribution to the predictive value.

For the Chilean dropout prediction, future work includes developing a model for predicting dropout within the same year. This was not actually possible with public data available since attendance and grades of students are not reported monthly. Another challenge is adapting the model for years 2020 and 2021 where students received mostly remote classes during COVID-19 pandemic. Attendance was measured differently (if students attended at least one online class at day) and curriculum was shortened and adapted to circumstances. Therefore, the continuity of measurement in attendance and school performance broke and they should be considered as additional and separated variables in the model. Additionally, the pandemic had an impact on socio-economic status because of parents' unemployment or death and until today there is not an actualized income data since SIMCE test and surveys were suspended in 2020 and 2021.

School failure was a diminishing problem, but the pandemic of COVID-19 will push the poorest students outside the system, especially women. Therefore, developing EWS systems with evidence-based strategies at school and territorial level should be carried out, to prevent children from abandoning their studies. The methodology proposed comprises the necessary steps to develop models with high predictive power if proper data is available.

We expect that the methodology and case presented in this article helps practitioners and public decision makers to create their own models to predict school failure, but also motivates them to capture, clean and systematize data to allow developing such kinds of systems.

Data availability

Specially thanks to the Center of Studies from the Ministry of Education, Agency of Quality of Education and JUNJI for providing special datasets to develop the model.

The data that support the findings of this study are available from Ministry of Education – Open data platform, but restrictions apply to the availability of these data, which were used under licence for the current study, and so are not publicly available. Data are however available from the authors upon reasonable request and with permission of Ministry of Education, Agency of Quality of Education and/or JUNJI.

Year where a student enrolls in first grade.

There is a fourth school category (delegated administration) where schools have a mechanism of funding by charters, with a basal funding to public property schools whose administration is delegated to private agents (Browne, 2017 ). Nevertheless, since there are only 70 schools in this category (41,578 students in 2019, 1.4% of total same year students) and notorious differences with respect to the ownership, funding, and administration of the schools, we decided to omit it from most of the reports in this article.

https://www.agenciaeducacion.cl/simce/

Adelman, M., Haimovich, F., Ham, A., & Vazquez, E. (2018). Predicting school dropout with administrative data: New evidence from Guatemala and Honduras. Education Economics, 26 (4), 356–372. https://doi.org/10.1080/09645292.2018.1433127

Article   Google Scholar  

Anderson, S., Uribe, M., & Valenzuela, J. P. (2021).Reforming public education in Chile: The creation of local education services. Educational Management Administration & Leadership , 1741143220983327. https://doi.org/10.1177/1741143220983327 .

Bentéjac, C., Csörgő, A., & Martínez-Muñoz, G. (2021). A comparative analysis of gradient boosting algorithms. Artificial Intelligence Review, 54 (3), 1937–1967. https://doi.org/10.1007/s10462-020-09896-5

Bergstra, J., Bardenet, R., Bengio, Y., & Kégl, B. (2011). Algorithms for Hyper-Parameter Optimization. Advances in Neural Information Processing Systems , 24 . https://papers.nips.cc/paper/2011/hash/86e8f7ab32cfd12577bc2619bc635690-Abstract.html .

Boniolo, P., & Najmias, C. (2018). School dropout and school lag in Argentina: A social classes approach. Tempo Social , 30 (3), 217–247. https://doi.org/10.11606/0103-2070.ts.2018.121349 .

Browne, M. (2017). Análisis del Sistema de Administración Delegada creada por el DL N o 3166 de 1980. Ministerio de Educación-SETP . http://biblioteca.digital.gob.cl/handle/123456789/897 . Accessed 20 Aug 2022.

Buenadicha, C., Galdon, G., Hermosilla, M., Loewe, D., & Pombo, C. (2019). La gestión ética de los datos . Inter-American Development Bank. https://doi.org/10.18235/0001623 .

Chen, T., & Guestrin, C. (2016). XGBoost: A Scalable Tree Boosting System. Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining , 785–794. https://doi.org/10.1145/2939672.2939785 .

Dos Santos, E. M., Sabourin, R., & Maupin, P. (2009). Overfitting cautious selection of classifier ensembles with genetic algorithms. Information Fusion, 10 (2), 150–162. https://doi.org/10.1016/j.inffus.2008.11.003

Dussaillant, F. (2017). Deserción escolar en Chile. Propuestas para la investigación y la política pública. Documento N o 18 , 1–18. Available at: https://gobierno.udd.cl/cpp/files/2020/10/18-Deserción.pdf . Accessed 20 Aug 2022.

Ecker-Lyster, M., & Niileksela, C. (2016). Keeping Students on Track to Graduate: A Synthesis of School Dropout Trends, Prevention, and Intervention Initiatives. The Journal of at-Risk Issues, 19 (2), 24–31.

Google Scholar  

Ester, M., Kriegel, H.-P., Sander, J., & Xu, X. (1996). A density-based algorithm for discovering clusters in large spatial databases with noise. Proceedings of the Second International Conference on Knowledge Discovery and Data Mining , 226–231. Available at: https://www.aaai.org/Papers/KDD/1996/KDD96-037.pdf .

Gil, A. J., Antelm-Lanzat, A. M., Cacheiro-González, M. L., & Pérez-Navío, E. (2019). School dropout factors: A teacher and school manager perspective. Educational Studies, 45 (6), 756–770. https://doi.org/10.1080/03055698.2018.1516632

Hirakawa, Y., & Taniguchi, K. (2021). School dropout in primary schools in rural Cambodia: School-level and student-level factors. Asia Pacific Journal of Education, 41 (3), 527–542. https://doi.org/10.1080/02188791.2020.1832042

Höfter, R. H. (2006). Private health insurance and utilization of health services in Chile. Applied Economics, 38 (4), 423–439. https://doi.org/10.1080/00036840500392797

Jaderberg, M., Dalibard, V., Osindero, S., Czarnecki, W. M., Donahue, J., Razavi, A., Vinyals, O., Green, T., Dunning, I., Simonyan, K., Fernando, C., & Kavukcuoglu, K. (2017). Population Based Training of Neural Networks. ArXiv:1711.09846 [Cs]. http://arxiv.org/abs/1711.09846.

Jena, M., & Dehuri, S. (2020). DecisionTree for Classification and Regression: A State-of-the Art Review. Informatica , 44 (4), 4. https://doi.org/10.31449/inf.v44i4.3023 .

Kattan, R. B., & Székely, M. (2017). Analyzing Upper Secondary Education Dropout in Latin America through a Cohort Approach. Journal of Education and Learning, 6 (4), 12–39. https://doi.org/10.5539/jel.v6n4p12

Ke, G., Meng, Q., Finley, T., Wang, T., Chen, W., Ma, W., Ye, Q., & Liu, T.-Y. (2017). LightGBM: a highly efficient gradient boosting decision tree. In Proceedings of the 31st International Conference on Neural Information Processing Systems (NIPS'17). Curran Associates Inc., Red Hook, NY, USA, 3149–3157

Khan, M. J., & Ahmed, J. (2021). Child education in the time of pandemic: Learning loss and dropout. Children and Youth Services Review, 127 , 106065. https://doi.org/10.1016/j.childyouth.2021.106065

Kursa, M. B., Jankowski, A., & Rudnicki, W. R. (2010). Boruta – A System for Feature Selection. Fundamenta Informaticae, 101 (4), 271–285. https://doi.org/10.3233/FI-2010-288

Article   MathSciNet   Google Scholar  

Ladd, H., & Fiske, E. (2020). International perspectives on school choice . Routledge.

Lee, S., & Chung, J. Y. (2019). The Machine Learning-Based Dropout Early Warning System for Improving the Performance of Dropout Prediction. Applied Sciences, 9 (15), 3093. https://doi.org/10.3390/app9153093

Lee-St John, T. J., Walsh, M. E., Raczek, A. E., Vuilleumier, C. E., Foley, C., Heberle, A., Sibley, E., & Dearing, E. (2018). The Long-Term Impact of Systemic Student Support in Elementary School: Reducing High School Dropout. Aera Open , 4 (4). https://doi.org/10.1177/2332858418799085 .

Levin, H. M., Belfield, C., Hollands, F., & Bowden, A. B. (2012). Cost-Effectiveness analysis of interventions that improve high school completion . Center for Benefit-Cost Studies of Education 34. https://repository.upenn.edu/cbcse/34 . Accessed 20 Aug 2022

Lundberg, S. M., Erion, G. G., & Lee, S.-I. (2019). Consistent Individualized Feature Attribution for Tree Ensembles. ArXiv:1802.03888 [Cs, Stat]. http://arxiv.org/abs/1802.03888.

Lundberg, S. M., Nair, B., Vavilala, M. S., Horibe, M., Eisses, M. J., Adams, T., Liston, D. E., Low, D.K.-W., Newman, S.-F., Kim, J., & Lee, S.-I. (2018). Explainable machine-learning predictions for the prevention of hypoxaemia during surgery. Nature Biomedical Engineering, 2 (10), 749–760. https://doi.org/10.1038/s41551-018-0304-0

Márquez-Vera, C., Cano, A., Romero, C., Noaman, A. Y. M., Fardoun, H. M., & Ventura, S. (2016). Early dropout prediction using data mining: A case study with high school students. Expert Systems, 33 (1), 107–124. https://doi.org/10.1111/exsy.12135

McInnes, L., Healy, J., & Melville, J. (2020). UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction ( arXiv:1802.03426 ). arXiv. https://doi.org/10.48550/arXiv.1802.03426 .

Mduma, N., Kalegele, K., & Machuve, D. (2019). A Survey of Machine Learning Approaches and Techniques for Student Dropout Prediction. Data Science Journal, 18 , 14. https://doi.org/10.5334/dsj-2019-014

Misra, P., & Yadav, A. (2020). Improving the classification accuracy using recursive feature elimination with cross-validation. International Journal on Emerging Technologies , 11 (3), 659-665.

Şara, N-B., Halland, R., Igel, C., and Alstrup, S. (2015). High-school dropout prediction using machine learning: a Danish large-scale study. In M. Verleysen (Ed.), Proceedings. ESANN 2015: 23rd European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning (pp. 319-324).

OECD. (2010). Overcoming school failure: Policies that work. OECD project description, (April). Available at https://www.oecd.org/education/school/45171670.pdf

OECD. (2020). Education at a Glance 2020: OECD Indicators . Organisation for Economic Co-operation and Development. https://www.oecd-ilibrary.org/education/education-at-a-glance-2020_69096873-en . Accessed 20 Aug 2022.

Pereira de Souza, C. M., Pereira, J. M., & de Jesus Ranke, M. da C. (2020). Reflexes of the Pandemic in school dropout/exit: The democratization of access and permanence. Revista Brasileira De Educacao Do Campo-Brazilian Journal of Rural Education , 5 , e10844. https://doi.org/10.20873/uft.rbec.e10844 .

Prokhorenkova, L., Gusev, G., Vorobev, A., Dorogush, A. V., & Gulin, A. (2019). CatBoost: Unbiased boosting with categorical features ( arXiv:1706.09516 ). arXiv. https://doi.org/10.48550/arXiv.1706.09516 .

Sahin, S., Arseven, Z., & Kilic, A. (2016). Causes of Student Absenteeism and School Dropouts. International Journal of Instruction , 9 (1), 195–210. https://doi.org/10.12973/iji.2016.9115a .

Sansone, D. (2019). Beyond Early Warning Indicators: High School Dropout and Machine Learning. Oxford Bulletin of Economics and Statistics, 81 (2), 456–485. https://doi.org/10.1111/obes.12277

Sharma, P., Mirzan, S. R., Bhandari, A., Pimpley, A., Eswaran, A., Srinivasan, S., & Shao, L. (2020). Evaluating Tree Explanation Methods for Anomaly Reasoning: A Case Study of SHAP TreeExplainer and TreeInterpreter. In G. Grossmann & S. Ram (Eds.), Advances in Conceptual Modeling (pp. 35–45). Springer International Publishing. https://doi.org/10.1007/978-3-030-65847-2_4 .

Sorensen, L. C. (2019). “Big Data” in Educational Administration: An Application for Predicting School Dropout Risk. Educational Administration Quarterly, 55 (3), 404–446. https://doi.org/10.1177/0013161X18799439

Studer, S., Bui, T. B., Drescher, C., Hanuschkin, A., Winkler, L., Peters, S., & Müller, K.-R. (2021). Towards CRISP-ML(Q): A Machine Learning Process Model with Quality Assurance Methodology. Machine Learning and Knowledge Extraction, 3 (2), 392–413. https://doi.org/10.3390/make3020020

UNESCO. (2012). International Standard Classification of Education ISCED 2011 . UNESCO Institute of Statistics, Montreal. Available at http://uis.unesco.org/sites/default/files/documents/international-standard-classification-of-education-isced-2011-en.pdf . Accessed 20 Aug 2022

UNESCO. (2020). UNESCO COVID-19 education response: How many students are at risk of not returning to school? Advocacy paper. UNESCO Paris. Available at https://unesdoc.unesco.org/ark:/48223/pf0000373992 . Accessed 20 Aug 2022.

Valenzuela, J. P., & Allende, C. (2014). Trayectorias de mejoramiento en el Sistema Escolar Chileno: Las escuelas de educación básica 2002 - 2010. Apuntes sobre Mejoramiento Escolar N°1, Enero 2014. Anillo de Ciencias Sociales sobre Mejoramiento de la Efectividad Escolar en Chile. https://www.mejoramientoescolar.cl/download.php?file=recursos/nota_tecnica.pdf . Accessed 20 Aug 2022.

Weybright, E. H., Caldwell, L. L., Wegner, L., & Smith, E. A. (2017). Predicting secondary school dropout among South African adolescents: A survival analysis approach. South African Journal of Education , 37 (2), 1–11. https://doi.org/10.15700/saje.v37n2a1353 .

Yoshida, S. (2020). Verification of Usefulness of SHAP values in Interpretation of Decision Tree Models . The Japanese Society for Artificial Intelligence. https://confit.atlas.jp/guide/event/jsai2020/subject/3E5-GS-2-04/detail . Accessed 20 Aug. 20022.

Zaff, J. F., Donlan, A., Gunning, A., Anderson, S. E., Mcdermott, E., & Sedaca, M. (2017). Factors that Promote High School Graduation: A Review of the Literature. Educational Psychology Review , 447–476. https://doi.org/10.1007/s10648-016-9363-5 .

Download references

We thank the support from ANID/PIA/Basal Funds for Centers of Excellence FB0003 and ANID-FONDEF IT17I0006 grants.

Author information

Authors and affiliations.

Institute of Education and Center for Advanced Research in Education, Universidad de Chile, Periodista José Carrasco Tapia 75, 8330014, Santiago, Región Metropolitana, Chile

Patricio Rodríguez & Juan Pablo Valenzuela

Center for Advanced Research in Education, Universidad de Chile, Santiago, Chile

Alexis Villanueva

Department of Informatics, Universidad Técnica Federico Santa María, Santiago, Chile

Lioubov Dombrovskaia

You can also search for this author in PubMed   Google Scholar

Corresponding author

Correspondence to Patricio Rodríguez .

Ethics declarations

Competing interests.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Additional information

Publisher's note.

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

Appendix B List of variables used and their significance in the final model

Rights and permissions.

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Reprints and permissions

About this article

Rodríguez, P., Villanueva, A., Dombrovskaia, L. et al. A methodology to design, develop, and evaluate machine learning models for predicting dropout in school systems: the case of Chile. Educ Inf Technol 28 , 10103–10149 (2023). https://doi.org/10.1007/s10639-022-11515-5

Download citation

Received : 22 August 2022

Accepted : 05 December 2022

Published : 23 January 2023

Issue Date : August 2023

DOI : https://doi.org/10.1007/s10639-022-11515-5

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

  • Data science applications in education
  • Machine learning
  • Learning analytics
  • Student retention
  • Find a journal
  • Publish with us
  • Track your research

Stanford University

Along with Stanford news and stories, show me:

  • Student information
  • Faculty/Staff information

We want to provide announcements, events, leadership messages and resources that are relevant to you. Your selection is stored in a browser cookie which you can remove at any time using “Clear all personalization” below.

Image credit: Claire Scully

New advances in technology are upending education, from the recent debut of new artificial intelligence (AI) chatbots like ChatGPT to the growing accessibility of virtual-reality tools that expand the boundaries of the classroom. For educators, at the heart of it all is the hope that every learner gets an equal chance to develop the skills they need to succeed. But that promise is not without its pitfalls.

“Technology is a game-changer for education – it offers the prospect of universal access to high-quality learning experiences, and it creates fundamentally new ways of teaching,” said Dan Schwartz, dean of Stanford Graduate School of Education (GSE), who is also a professor of educational technology at the GSE and faculty director of the Stanford Accelerator for Learning . “But there are a lot of ways we teach that aren’t great, and a big fear with AI in particular is that we just get more efficient at teaching badly. This is a moment to pay attention, to do things differently.”

For K-12 schools, this year also marks the end of the Elementary and Secondary School Emergency Relief (ESSER) funding program, which has provided pandemic recovery funds that many districts used to invest in educational software and systems. With these funds running out in September 2024, schools are trying to determine their best use of technology as they face the prospect of diminishing resources.

Here, Schwartz and other Stanford education scholars weigh in on some of the technology trends taking center stage in the classroom this year.

AI in the classroom

In 2023, the big story in technology and education was generative AI, following the introduction of ChatGPT and other chatbots that produce text seemingly written by a human in response to a question or prompt. Educators immediately worried that students would use the chatbot to cheat by trying to pass its writing off as their own. As schools move to adopt policies around students’ use of the tool, many are also beginning to explore potential opportunities – for example, to generate reading assignments or coach students during the writing process.

AI can also help automate tasks like grading and lesson planning, freeing teachers to do the human work that drew them into the profession in the first place, said Victor Lee, an associate professor at the GSE and faculty lead for the AI + Education initiative at the Stanford Accelerator for Learning. “I’m heartened to see some movement toward creating AI tools that make teachers’ lives better – not to replace them, but to give them the time to do the work that only teachers are able to do,” he said. “I hope to see more on that front.”

He also emphasized the need to teach students now to begin questioning and critiquing the development and use of AI. “AI is not going away,” said Lee, who is also director of CRAFT (Classroom-Ready Resources about AI for Teaching), which provides free resources to help teach AI literacy to high school students across subject areas. “We need to teach students how to understand and think critically about this technology.”

Immersive environments

The use of immersive technologies like augmented reality, virtual reality, and mixed reality is also expected to surge in the classroom, especially as new high-profile devices integrating these realities hit the marketplace in 2024.

The educational possibilities now go beyond putting on a headset and experiencing life in a distant location. With new technologies, students can create their own local interactive 360-degree scenarios, using just a cell phone or inexpensive camera and simple online tools.

“This is an area that’s really going to explode over the next couple of years,” said Kristen Pilner Blair, director of research for the Digital Learning initiative at the Stanford Accelerator for Learning, which runs a program exploring the use of virtual field trips to promote learning. “Students can learn about the effects of climate change, say, by virtually experiencing the impact on a particular environment. But they can also become creators, documenting and sharing immersive media that shows the effects where they live.”

Integrating AI into virtual simulations could also soon take the experience to another level, Schwartz said. “If your VR experience brings me to a redwood tree, you could have a window pop up that allows me to ask questions about the tree, and AI can deliver the answers.”

Gamification

Another trend expected to intensify this year is the gamification of learning activities, often featuring dynamic videos with interactive elements to engage and hold students’ attention.

“Gamification is a good motivator, because one key aspect is reward, which is very powerful,” said Schwartz. The downside? Rewards are specific to the activity at hand, which may not extend to learning more generally. “If I get rewarded for doing math in a space-age video game, it doesn’t mean I’m going to be motivated to do math anywhere else.”

Gamification sometimes tries to make “chocolate-covered broccoli,” Schwartz said, by adding art and rewards to make speeded response tasks involving single-answer, factual questions more fun. He hopes to see more creative play patterns that give students points for rethinking an approach or adapting their strategy, rather than only rewarding them for quickly producing a correct response.

Data-gathering and analysis

The growing use of technology in schools is producing massive amounts of data on students’ activities in the classroom and online. “We’re now able to capture moment-to-moment data, every keystroke a kid makes,” said Schwartz – data that can reveal areas of struggle and different learning opportunities, from solving a math problem to approaching a writing assignment.

But outside of research settings, he said, that type of granular data – now owned by tech companies – is more likely used to refine the design of the software than to provide teachers with actionable information.

The promise of personalized learning is being able to generate content aligned with students’ interests and skill levels, and making lessons more accessible for multilingual learners and students with disabilities. Realizing that promise requires that educators can make sense of the data that’s being collected, said Schwartz – and while advances in AI are making it easier to identify patterns and findings, the data also needs to be in a system and form educators can access and analyze for decision-making. Developing a usable infrastructure for that data, Schwartz said, is an important next step.

With the accumulation of student data comes privacy concerns: How is the data being collected? Are there regulations or guidelines around its use in decision-making? What steps are being taken to prevent unauthorized access? In 2023 K-12 schools experienced a rise in cyberattacks, underscoring the need to implement strong systems to safeguard student data.

Technology is “requiring people to check their assumptions about education,” said Schwartz, noting that AI in particular is very efficient at replicating biases and automating the way things have been done in the past, including poor models of instruction. “But it’s also opening up new possibilities for students producing material, and for being able to identify children who are not average so we can customize toward them. It’s an opportunity to think of entirely new ways of teaching – this is the path I hope to see.”

University of Wyoming Logo

  • Apply to UW
  • Programs & Majors
  • Cost & Financial Aid
  • Current Students
  • UW Libraries
  • Degree Plans & Courses
  • Advising & Career Services
  • UW College of Law
  • Honors College
  • Academic Affairs
  • Geological Museum
  • All Colleges
  • Campus Recreation
  • Campus Maps
  • Housing & Dining
  • Transit & Parking
  • University Store
  • Student Organizations
  • Campus Activities
  • Campus Safety
  • Diversity, Equity & Inclusion
  • Research & Economic Dev.
  • Wyoming INBRE
  • Neuroscience Center
  • Technology Business Center
  • National Parks Service
  • Research Production Center
  • Supercomputing
  • Water Research
  • WY EPSCoR/IDeA
  • American Heritage Center
  • Where We Shine
  • About Laramie
  • Student Stories
  • Campus Fact Book
  • UWYO Magazine
  • Marketing & Brand Center
  • Administrative Resources
  • Strategic Plan
  • +Application Login
  • UW Homepage

UW’s SER and College of Engineering and Physical Sciences Fund Energy-Related Equipment Proposals

  • News Releases
  • Back to 2023 Archive

Institutional Communications Bureau of Mines Building, Room 137 Laramie, WY 82071 Phone: (307) 766-2929 Email:   [email protected]

Find us on Facebook (Link opens a new window)

Published May 15, 2024

Recipients have been selected from among 41 proposals for funding following a request from the University of Wyoming’s School of Energy Resources (SER) to UW employees in need of energy-related laboratory equipment -- including repairs -- or supplies to conduct energy-related research.

“The volume of proposals, totaling more than 10 times the funds SER had available, was extremely telling of the need for support for equipment to bolster the UW energy research enterprise,” says SER Executive Director Holly Krutka. “It quickly became apparent that we needed to find a way to support more proposals than we had initially expected, and we’re grateful to our partners in the College of Engineering and Physical Sciences for stepping in.”

Recognizing the need and merit among the proposals, the College of Engineering and Physical Sciences joined the commitment to fund additional equipment proposals from multiple departments around campus.

“We want to set up our researchers for success, facilitate the shared access to new and existing equipment, and promote further collaboration across UW colleges and departments,” says Cameron Wright, the Carrell Family Dean of the College of Engineering and Physical Sciences. “We’re grateful for our increased Tier 1 funding, which allowed us to make these investments to improve our research tools and facilities. We look forward to seeing these new purchases and upgrades put to good use.”

Proposals were considered for repairing or upgrading existing equipment up to $100,000; new equipment purchases up to $500,000; and requests for purchasing supplies, such as laboratory and testing supplies or computational resources, up to $100,000.

Eight proposals were funded by SER, and nine were funded by the College of Engineering and Physical Sciences, with a total of $4.5 million allocated for equipment purchases.

Selected recipients are:

-- Kyle Summerfield, program manager of SER’s 3D Visualization Center, and Jeff Hamerlinck, director of the Wyoming Geographic Information Science Center, were selected for their requested high-resolution data visualization LED wall system along with matching funds from the School of Computing.

-- A proposal by John Hoberg, a professor in the Department of Chemistry, was selected for a high-pressure hydrogenation reactor.

-- Haibo Zhai, the Roy and Caryl Cline Distinguished Chair and a professor in the Department of Civil and Architectural Engineering and Construction Management, will receive funding for laboratory computers.

-- Jonathan Brant, a professor in the Department of Civil and Architectural Engineering and Construction Management and SER’s Center of Excellence in Produced Water Management, will receive funding for an integrated electrolysis system for hydrogen production.

-- Shane Murphy, an associate professor in the Department of Atmospheric Science and SER’s Center for Air Quality, will receive funding for repairs and upgrades to existing equipment in the Atmospheric Science Mobile Lab.

-- A collaborative proposal from Fabian Nippgen and Kristina Hufford, both associate professors in the Department of Ecosystem Science and Management, along with Roger Coupal, a professor in the Department of Agricultural and Applied Economics, will receive funding for a fixed-wing unmanned aerial vehicle and associated sensor to measure discharge and trigger water sampling for reclaimed mine land.

-- Grant Copeland, an SER research professional in the Center for Economic Geology Research, submitted a successful proposal for an ashing furnace and hood to improve current critical mineral research capabilities.

-- Jonathan Naughton, a professor in the Department of Mechanical Engineering and director of the Wind Energy Research Center, will receive funding for a laser-particle image velocity system.

-- Minou Rabiei, an associate professor in the Department of Energy and Petroleum Engineering, will receive funding for a rock compressibility system for the testing and characterization of cavern storage formations for carbon capture and storage.

-- Tawfik Elshehabi, an assistant professor in the Department of Energy and Petroleum Engineering, will receive funding for Eco-Cem liquid cement (slurry) testing equipment.

-- Kam Ng, a professor in the Department of Civil and Architectural Engineering and Construction Management, submitted a successful proposal for a 4D X-ray microtomography system.

-- Vamegh Rasouli, a professor and head of the Department of Energy and Petroleum Engineering, will receive funding for a G200X nanoindenter from KLA Instruments, which will be used for nanoscale mechanical testing.

-- Soheil Saraji, an associate professor in the Department of Energy and Petroleum Engineering, will receive funding for a micro-X-ray fluorescence spectrometer and an automated petrography polarization microscope for geological formation characterization capabilities.

-- Maddie Lewis, an assistant professor in the Department of Geology and Geophysics, will receive funding for instrument, electrical and venting renovations in a lab dedicated to the characterization of critical minerals.

-- Brian Leonard, a professor and head of the Department of Chemistry, will receive funding for four new furnaces, ceramic reaction tubes, thermocouples and heating elements.

-- Brad Carr, an associate research scientist in the Department of Geology and Geophysics, will receive funding for a SmartSolo seismic receiver, a Geometrics Stratagem EH-5 magnetotelluric system and a Proceq Screening Eagle multifrequency ground penetrating radar system for drill pad technology.

-- David Reid, a graduate teaching assistant, and Rasouli, both in the Department of Energy and Petroleum Engineering, submitted a successful request for a HPHT foam rheometer for the Integrated Research Lab.

Overall, the proposals will support individuals in 14 departments and centers of excellence on a range of energy-related projects.

“We are very pleased to see the desire of our researchers and faculty to do so much more than they are already doing to drive quality energy exploration,” Wright says. “Providing access to proper tools and equipment will make a huge difference in the outcomes for those who are already dedicated to making a tangible difference in Wyoming’s energy industry.”

Kathleen R. Bogart Ph.D.

Psych Careers

A graduation message for psychology students, personal perspective: lessons on purpose, persistence, and community..

Posted May 9, 2024 | Reviewed by Monica Vilhauer

Last year, I delivered the graduation speech for Oregon State University's School of Psychological Science, and I think the message is even more relevant today.

Three lessons on purpose , persistence , and community can help you navigate life's next chapter. From psychological research, we know that facts and figures are less impactful than personal stories, so I’ll pepper in examples from my own experiences along the way.

Think back to why you decided to become a psychology major. Was it because of a desire to help people? An insatiable curiosity about human nature? To make a difference in people’s lives? For me, it was all of the above.

Kathleen Bogart

I have been interested in psychology ever since I was born. I was born with Moebius syndrome , a disability characterized by facial paralysis and the inability to move my eyes from side to side. At an early age, I understood that the way I communicated was unusual, that people were confused by my lack of facial expression. I became fascinated with communication and social interaction.

These interests led me to study psychology as a college student. Toward the end of my bachelor’s I set out to do my very first college term paper on Moebius syndrome. I showed up at the library expecting to find pages and pages of answers, but I discovered there was only a handful of psych papers published on it! This was bad news for two reasons: First, I didn’t have enough sources to write my term paper. Second, my chosen field had not included people like me.

I realized I was at a crossroads. I could give up and choose another path, or I could start developing the psychological knowledge in this area.

I chose the latter. I knew that I had the unique motivation and insight to grow this field. So I applied to graduate school—but the first time I applied, I was rejected from every single program.

Graduate training in psychology research follows a mentorship model, and because there were few psychologists studying disability, and few psychologists who had disabilities themselves, I struggled to find an advisor who was interested in this topic. Eventually, I found supportive allies to be my mentors. I was the speaker at my own PhD graduation ceremony, when my mentor Dr. Linda Tickle-Degnen hooded me. Just over 10 years later, I spoke at the graduation ceremony, where I am now a faculty member, and hooded my first disabled PhD student.

I've spent more than 15 years studying ableism, or prejudice toward people with disability. Nearly 20% of Americans have a disability, making it one of the largest minority groups in the U.S. And one that is now a little less underrepresented in psychology.

My experience made me acutely aware of the importance of finding purpose to live a fulfilling life. Personally, my work provides meaning by helping others with similar conditions and teaching students about a broader and more diverse swath of humanity.

I encourage you to find meaning in your work. It doesn’t have to be as entrenched in your identity as mine, and it doesn’t even have to be connected to you job. But find a field, a project, or a hobby in which you feel an intense curiosity, an excitement for learning, a passion for change, and it will drive you to persevere. Success will follow. Studies consistently show that individuals who find meaning and purpose in their work are more engaged, fulfilled, and resilient.

Purpose is your own personal mission statement. What is yours? It could be to love your fellow humans, it could be to help others. Prioritize actions that align with your mission.

Mine is to make the world a more inclusive place, using psychology!

Persistence

The quote from Thomas Edison that "genius is one percent inspiration and 99 percent perspiration" rings true to me. My successes are due to simple perseverance, as well as a lot of support (see lesson on community below). What keeps me going is that I find great personal meaning in my work.

My day-to-day work is not glamorous. I spend my time working with students or sitting alone in a room and writing. I make a commitment to write around the same time every day. Disabled advocate Cassie Winter calls this type of work “butt in chair time.” This simply means creating a consistent schedule to work on your priorities. Sometimes this means staring blankly and thinking through ideas; other times it means writing furiously in a flow state. My butt in chair time creates a sustainable pace, instead of falling into boom or bust cycles, and prevents burnout .

research proposal school dropout

Your work and hobbies may look different from mine. Swap butt in chair time to boots on the ground time, or whatever resonates with you. The point is, prioritize time to work on the things that matter to you.

Research links persistence with a growth mindset . It is important to note that the healthy kind of persistence involves flexibility, not ridged stubbornness. Albert Einstein said it well when he said “insanity is doing the same thing over and over and expecting different results.” Failure is feedback that we can learn and grow from. Change your approach and try again.

For example, when I didn’t get into any graduate schools the first time around, I realized I needed to change my strategy. I studied for the GRE using a different approach and retook it. I also broadened my search to other areas of the country and applied to masters programs. I first got into a masters program, which gave me the opportunity to hone my skills. Then I was ready to move into my goal, a PhD program.

In college and graduate school, I ached for friends and role models who identified as disabled but found none. My experience made me acutely aware of the need for better representation of marginalized people in higher education .

Now, I teach a class at OSU on the Psychology of Disability, where I give students with and without disabilities an opportunity to see representation of this important minority group and its intersections. I also co-founded the Disability Advocacy Research Network ( DARN ), an organization for disabled psychologists and students to find the community that I didn’t have earlier in my career . Last year, I delivered the speech at OSU's first disability graduation ceremony. I am so heartened that the next generation will be better able to find community.

That brings me to my last piece of advice. Find your community, or create it: a place where you can be authentically you.

As you enter a new stage in life, community will become all the more important. Seek out mentors who can guide you. Likewise, you are now in a position where you could mentor people who are just entering college. Studies have shown that strong social connections contribute to resilience and overall life satisfaction. In creating them, we not only enrich our own lives but also create a ripple effect of support in the lives of those around us.

Kathleen R. Bogart Ph.D.

Kathleen Bogart, Ph.D. , is an Associate Professor of Psychology at Oregon State University. She researches the implications of living with disability, rare disorders, or facial differences.

  • Find a Therapist
  • Find a Treatment Center
  • Find a Psychiatrist
  • Find a Support Group
  • Find Online Therapy
  • United States
  • Brooklyn, NY
  • Chicago, IL
  • Houston, TX
  • Los Angeles, CA
  • New York, NY
  • Portland, OR
  • San Diego, CA
  • San Francisco, CA
  • Seattle, WA
  • Washington, DC
  • Asperger's
  • Bipolar Disorder
  • Chronic Pain
  • Eating Disorders
  • Passive Aggression
  • Personality
  • Goal Setting
  • Positive Psychology
  • Stopping Smoking
  • Low Sexual Desire
  • Relationships
  • Child Development
  • Therapy Center NEW
  • Diagnosis Dictionary
  • Types of Therapy

May 2024 magazine cover

At any moment, someone’s aggravating behavior or our own bad luck can set us off on an emotional spiral that threatens to derail our entire day. Here’s how we can face our triggers with less reactivity so that we can get on with our lives.

  • Emotional Intelligence
  • Gaslighting
  • Affective Forecasting
  • Neuroscience

Steve Hicks School of Social Work

SHS Certifies Four Doctoral Candidates for Graduation in the Class of 2024

research proposal school dropout

On Saturday, May 11, Dean Allan Cole certified the graduation of four Ph.D. candidates, signifying the newest generation of doctors who will impact social work and social work education. SHS is proud to welcome these four individuals into the echelon of SHS graduates:

ANA CHATHAM, LCSW

research proposal school dropout

Dr. Ana Chatham , a licensed clinical social worker in the state of Texas, focused her research at SHS on improving the mental health and wellbeing of Latinx persons through culturally informed prevention, intervention, and structural changes. Utilizing her expertise in community-based participatory research (CBPR), her research is grounded in her 10 years of practice experience serving clients in micro, mezzo, and macro capacities in fields including family preservation, domestic violence, and mental health.

Mentored by Dr. Carmen Valdez , Dr. Chatham is a collaborator in various interdisciplinary participatory studies including a mental health needs assessment of rural communities in Mexico and a study on minoritized youth’s experiences of environmental justice. Ana is a licensed clinical social worker in Texas who graduated from Seton Hill University with a B.A. in Psychology and from Baylor University with an MSW.

Her dissertation, entitled Factors Affecting the Mental Health of Youth in Rural Puebla, Mexico: A Photovoice Project , was supervised by Carmen Valdez, Ph.D.

“ I am thankful for the mentoring and opportunities I received at SHS, through which I have been able to grow as a scholar,” said Dr. Chatham. “Conducting community-based research focusing on the mental health needs and experiences of Latino youth both domestically and abroad alongside Dr. Valdez has been intellectually stimulating and emotionally rewarding. I am filled with gratitude and thrilled for what is ahead. “

research proposal school dropout

Dr. Xiao Ding , the recipient of the Michael R. Daley Endowed Presidential Scholarship for Doctoral Students, focused her research on the impact of intervening in ecological systems, such as parent-child relationships or school culture and environment, using solution-focused brief therapy (SFBT) and analyzing students’ mental health/behavioral health and academic outcomes.

During her Ph.D. training, Xiao gained knowledge and expertise in SFBT by continuously practicing as an LMSW and serving as an outside consultation staff member with Dr. Cynthia Franklin at the Gonzalo Garza Independence High School, a model solution-focused alternative school for dropout prevention. In addition, Xiao has been involved with several systematic reviews and meta-analyses on SFBT and is conducting a meta-analysis on interventions at independent alternative high school settings.

Xiao earned her Master of Science in Social Administration (MSSA; equivalent to an MSW) from Case Western Reserve University in 2018, with a concentration in Mental Health, Children, Youth, and Families. She also holds a Bachelor of Arts degree in Broadcasting and Television Journalism from Hubei University, China. Born and raised in an underdeveloped, remote, and ethnic minority-dominant city in the northwest of mainland China, Xiao witnessed the regional disparities among ethnic and socioeconomic groups regarding access to social services and the need for change.

Her dissertation, entitled Promoting Mental Health and Well-being in Alternative High Schools: A School-Based Solution-Focused Approach to Crisis Prevention and Early Intervention , was supervised by Cynthia Franklin, Ph.D.

“Ph.D. study is a journey – travel it well,” said Dr. Ding. “I am honored and privileged to have been guided by admirable mentors and supported by trustworthy companions throughout this journey. My time at SHS and UT Austin has profoundly shaped me both as a human being and as a scholar. These years have become an irreplaceable chapter in my life.”

LAURA HARJEET DOSANJH

research proposal school dropout

Dr. Laura Harjeet Dosanjh , is a recipient of the Michael R. Daley Endowed Presidential Scholarship for Doctoral Students and an awardee of the Provost’s Early Career Fellowship award. Her research at SHS has been focused on the impact of Adverse Childhood Experiences (ACEs) on global life outcomes including the physiological stress response. 

During her studies, Dr. Dosanjh has collaborated on interdisciplinary research focusing on the impact of ACEs in minoritized communities and families. Her collaborators have included the Texas Institute for Child and Family Wellbeing, the Thriving Together research project with Dr. Esther Calzada , and smaller projects with Dr. Catherine Cubbin and Dr. Frances Champagne . She plans to continue building knowledge in the field of social work through research on ACEs and associated outcomes, the neurobiology of trauma, and factors that protect against the effects of chronic stress. Her commitment is to focus her life’s work on these issues to support better physical and mental health outcomes for all people, with a particular emphasis on health disparities in sexual and gender minority communities.

Laura received a Bachelor of Arts degree from the University of British Columbia, and a Masters of Arts in Expressive Arts Therapy, minor Psychology from the European Graduate School. Her current research is informed by almost a decade of clinical experience working on the front lines with adult clients experiencing complex concurrent disorders. 

Her dissertation, entitled  The Impact of Inflammation and Minority Stress in a Moderated Mediation Model Examining Childhood Adversity and Psychological Distress in Young Men with Sexually Diverse Identity, was supervised by  Cynthia Franklin, Ph.D.

“My experience as a doctoral student at the SHS has been incredible,” said Dr. Dosanjh. “I have had the privilege of learning from leading national scholars whose guidance and expertise have nurtured both my personal and professional growth. I feel very lucky to be part of an academic community that is dedicated to empirical excellence, social justice, and everyday kindness. The sense of belonging and support I have found here has been truly heartwarming. I look forward to applying all of these values to my work as an NIH-funded post-doctoral fellow at UT’s Population Research Center in the fall, carrying forward the commitment to equity and advocacy that SHSSW embodies.”

KATIE MCCORMICK

Katie McCormick

Dr. Katie McCormick , a recipient of the Donald D. Harrington Dissertation Fellowship, focused her research at SHS on multi-level, community-centered approaches to supporting addressing the intersecting HIV and opioid epidemics. Katie has expertise collaboratively developing, implementing, and evaluating community-driven learning collaboratives and training community-based organizations in adopting harm reduction approaches.

Dr. McCormick received her Bachelor’s in Social Work from Baylor University and her Master’s in Social Work from the University of Houston. Her professional interests include harm reduction, implementation science, community-based participatory research, and multi-level interventions.

Her dissertation, entitled Occupational Stress and Burnout Among Texas Harm Reduction Workers , was supervised by Lori K. Holleran Steiker, Ph.D. and Kasey R. Claborn, Ph.D.

“My time at SHS has been better than I imagined,” said Dr. McCormick. “Our accomplished faculty are deeply invested in students’ personal and professional development, and enjoy empowering students to take charge of their own learning and research agendas. SHS has afforded me access to countless opportunities throughout and beyond UT’s campus that have enabled me to translate my research to the field and facilitate real-world impact to achieve social justice. I am proud to be a UT SHS graduate, and ecstatic to continue my time at UT through a NIDA-funded post-doc with Dell Medical School and SHS.”

On Saturday, May 10, Dean Allan Cole certified the graduation of these four doctors, along with more than 200 BSW and MSSW candidates in the Steve Hicks School of Social Work during 2024 commencement ceremonies.

Take the Quiz: Find the Best State for You »

What's the best state for you ».

Hamas Chief Says Israeli Amendments on Ceasefire Proposal Led to Deadlock

Reuters

Palestinian group Hamas' top leader, Ismail Haniyeh meets with Iranian President Ebrahim Raisi (not pictured) in Tehran, Iran March 27, 2024. Iran's Presidency/WANA (West Asia News Agency)/Handout via REUTERS/File Photo

By Nidal al-Mughrabi

CAIRO (Reuters) - Hamas chief Ismail Haniyeh blamed Israel on Wednesday for a deadlock in Gaza ceasefire negotiations and reiterated key demands including that any agreement provide a framework for a permanent end to Israel's offensive in the enclave.

Earlier this month, the Palestinian group said it agreed to a truce proposal from Qatari and Egyptian mediators that Israel previously accepted.

Israel disputed this, saying the three-phase proposal

War in Israel and Gaza

Palestinians are mourning by the bodies of relatives who were killed in an Israeli bombardment, at the al-Aqsa hospital in Deir Balah in the central Gaza Strip, on April 28, 2024, amid the ongoing conflict between Israel and the militant group Hamas. (Photo by Majdi Fathi/NurPhoto via Getty Images)

approved by Hamas was unacceptable because terms had been watered down. Ceasefire talks in Cairo broke up this month with no agreement.

"They also introduced amendments to the proposal that put the negotiation into a deadlock," said Haniyeh, who is based in Qatar, in a televised speech.

Haniyeh said his group was determined to pursue all available ways to end the war in Gaza, leaving the door open to more mediation efforts, but he held to the group's key demands.

"Any efforts or agreement must secure a permanent ceasefire, a comprehensive pull-out from all of the Gaza Strip, a real prisoner swap deal, the return of the displaced, reconstruction, and lifting the blockade," said Haniyeh.

Israel says it wants to reach a prisoners-for-hostages deal but has so far refused any commitment to end the military offensive in Gaza, which it said seeks to annihilate Hamas.

Haniyeh, whose group has been running Gaza since 2007, rejected any post-war settlement in Gaza that excludes the group. "Hamas existed to stay," he said.

"The movement (Hamas) will decide, along with all national factions, the administration of the Gaza Strip after the war," said Haniyeh.

Israel says Hamas can't have any role in ruling Gaza after the war is over. Its ally, the United States, says it wants to see Gaza and the West Bank reunited under Hamas' rival, the Palestinian Authority, which currently has limited self-rule in the Israeli-occupied West Bank.

Health authorities in Gaza say the war has killed almost 35,000 people since the Oct. 7 attack by Hamas on Israel. Some 1,200 people were killed in Israel and 253 were taken hostage on Oct. 7, according to Israeli tallies.

(Reporting by Nidal al-Mughrabi, additional reporting by Muhammad Al Gebaly; Editing by Chris Reese and Cynthia Osterman)

Copyright 2024 Thomson Reuters .

Join the Conversation

Tags: Israel , Middle East , Egypt

America 2024

research proposal school dropout

Health News Bulletin

Stay informed on the latest news on health and COVID-19 from the editors at U.S. News & World Report.

Sign in to manage your newsletters »

Sign up to receive the latest updates from U.S News & World Report and our trusted partners and sponsors. By clicking submit, you are agreeing to our Terms and Conditions & Privacy Policy .

You May Also Like

The 10 worst presidents.

U.S. News Staff Feb. 23, 2024

research proposal school dropout

Cartoons on President Donald Trump

Feb. 1, 2017, at 1:24 p.m.

research proposal school dropout

Photos: Obama Behind the Scenes

April 8, 2022

research proposal school dropout

Photos: Who Supports Joe Biden?

March 11, 2020

research proposal school dropout

Who Is Prime Minister Robert Fico?

Laura Mannweiler May 15, 2024

research proposal school dropout

Biden and Trump Agree to Debate

Lauren Camera May 15, 2024

research proposal school dropout

Biden Muddies Message on Israel

Aneeta Mathur-Ashton May 15, 2024

research proposal school dropout

Consumers Get a Break From Inflation

Tim Smart May 15, 2024

research proposal school dropout

Trump Team Gets Its Shot at Cohen

Lauren Camera May 14, 2024

research proposal school dropout

New China Tariffs: What to Know

Cecelia Smith-Schoenwalder May 14, 2024

research proposal school dropout

  • Skip to Content
  • Skip to Main Navigation
  • Skip to Search

research proposal school dropout

IUPUI IUPUI IUPUI

Open Search

  • Meet Dean Bravo
  • Mission Statement
  • Inlow Hall Code of Conduct
  • Inlow Hall Access Policy
  • Emergency Action Plan (EAP)
  • Evacuation Information
  • Emergency Preparedness
  • Welcome Letter
  • Pipeline Initiatives
  • Commitment to Diversity and Inclusion
  • Chairs & Professorships
  • Faculty Research
  • Our History
  • Recognition & Accolades
  • McKinney by the Numbers
  • Consumer Information (ABA Required Disclosures)
  • JD Professional Licensure Disclosure
  • About Indianapolis
  • Inlow Hall Virtual Tour via Google
  • Early Decision
  • Direct Admission
  • BA JD Applicants
  • Transfer Students
  • Visiting Students
  • International Students
  • Degree Requirements
  • Learning Outcomes
  • Part-time Hybrid
  • Graduate Certificates
  • Online Options
  • Dual Degrees
  • Frequently Asked Questions
  • Tuition & Fees
  • Scholarships
  • Financial Aid
  • COA Increase Request Form
  • Non-Resident Tuition Discounts
  • Partner School Awards
  • Bar Study Loan
  • Summer Law and Leadership Academy
  • Class Profiles
  • On Demand Videos
  • Information Sessions
  • J.D. Viewbook
  • Virtual Viewbook
  • Nuts & Bolts
  • Obligations
  • Admitted Students Events
  • Check Application Status
  • Program Tracks
  • LL.M. Courses
  • Application Requirements
  • Financial Information
  • FAQ / Community Forum
  • Practical Work Experience
  • Virtual Visits
  • S.J.D. Candidates
  • Preparing Your Application
  • M.J. Courses
  • M.J. Dual Degree Options
  • Financing Your Education
  • M.J. Contact Form
  • M.J. New Student Orientation
  • Admissions Workshops
  • LL.M. Virtual Chats & Open Houses
  • M.J. Virtual Information Sessions
  • Maps & Directions
  • ABA Required Disclosures
  • Request Info
  • McKinney Law Online
  • Required Courses (JD)
  • Experiential Credit
  • Elective Courses
  • Externships
  • Law Reviews
  • Legal Writing
  • Online Courses
  • Simulation Courses
  • Study Abroad
  • Courses for J.D. Graduate Certificates
  • Academic Calendar
  • Class & Exam Schedules
  • Watch & Learn (Videos)
  • Clinic Faculty & Staff
  • Rural Justice
  • Trial Advocacy Competition
  • Advanced Course-Related Experience (ACRE)
  • Advanced Field Research (AFR)
  • Legal Writing Program
  • Advocate FAQ
  • Sample Score Sheet
  • Executive Board
  • Order of the Barristers
  • Moot Court Faculty & Staff
  • Donate to Moot Court
  • Pro Bono Fair
  • Professional Conduct
  • Student Hours Log Sheet
  • Contact the Pro Bono Program
  • Skills Assessment
  • Cover Letter
  • Job Interview Tips
  • Thank You Letters
  • McKinney Works (Link)
  • Student Preparation
  • Employer Information
  • Professional Organizations
  • Judicial Clerkships Guide
  • Where Our Graduates Work
  • Loan Repayment Assistance Program (LRAP)
  • ABA Employment Summary Reports
  • Pro Bono Program
  • Environmental Law
  • Human Rights Law
  • Intellectual Property Law
  • International & Comparative Law
  • Joint Center for Asian Law Studies
  • Full Program Details
  • International Law News
  • Slavery Past, Present, Future: 5th Global Meeting
  • State Government Law
  • All Faculty & Staff
  • Administration
  • Administrative Staff
  • Faculty Assistants
  • Adjunct Faculty
  • Emeritus Faculty
  • Library Faculty
  • Library Staff
  • Media Experts
  • Hours and Access
  • Library Maps
  • Staff Directory
  • Recommended Reading
  • Reserve Request
  • Research Tools
  • Wellness & Health
  • News & Events
  • Alumni Association Board of Directors
  • Board of Visitors
  • Black Alumni Council
  • Dean's Global Alumni Council
  • Class Notes
  • 45th Class Reunion
  • 50th Class Reunion
  • Volunteer Opportunities
  • Nomination Form
  • Alumni Magazine
  • Continuing Legal Education (CLE)
  • Keep in Touch!
  • McKinney Law Fund
  • Law Firm Campaign
  • Class Agents
  • How to Create a Scholarship
  • Scholarships & Awards List
  • Ways to Give
  • Donor Recognition
  • Words of Gratitude
  • Office of Student Affairs
  • Commencement
  • Student Handbook
  • Online Forms
  • Bar Exam Info
  • Financing Your Bar Study
  • ABA and Bar Passage Information
  • Returning J.D. Student Scholarship Application
  • Scholarship Descriptions
  • Student Organizations
  • Event Policy
  • Student Wellness & Mental Health
  • Transcripts (Official)
  • Technology Requirement
  • Law School Intranet (login required)

IU Robert H. McKinney School of Law

  • Current Students
  • Law School News Story

IU McKinney Honors Graduating Students at Recognition Celebration

Jimmie McMillian

The law school’s graduation recognition event took place at the Murat Theater at Old National Centre on May 8. The law school celebrated 261 graduates from 17 countries, 25 states, and 31 Indiana counties. A total of 58% of the class is female, 42% is male, and 20% are people of color. The oldest graduate just shy of 64 and the youngest just turned 23.

Graduating students and their families heard remarks from Jimmie L. McMillian , J.D. ’02, senior corporate counsel and chief diversity officer of Penske Entertainment, a subsidiary of Penske Corporation. In those capacities, he provides legal counsel to the Indianapolis Motor Speedway, INDYCAR, and IMS Productions, and directs all corporate diversity, equity, and inclusion initiatives. His counsel responsibilities include negotiating the Speedway’s licensing, music concert, construction, and indemnity contracts. His diversity portfolio includes Penske’s “Race for Equality & Change,” an initiative to provide opportunities in the motorsports industry and open wheel racing for women and men from historically underrepresented groups.

Prior to his current assignment at Penske Entertainment, McMillian was a partner in the litigation department of Barnes & Thornburg, one of the nation’s 100 largest law firms. There he represented corporate clients in a wide range of litigation involving commercial contracts, products liability, personal injury and wrongful death claims, construction disputes, and internal and criminal investigations. After graduating from McKinney Law and before beginning work at Barnes & Thornburg, McMillian was a law clerk for former Indiana Supreme Court Justice Frank Sullivan, Jr., who is now a professor of business and commercial law courses at McKinney.

It wasn’t the first time McMillian was at the podium during such an event. He was the speaker for his graduating class in 2002 and he addressed the class of 2024 during their orientation exercises in 2021.

Student speakers included Makayla Phillips, who spoke on behalf of the M.J. division; Kenyi Kenyi, who spoke for the LL.M. division; Hunter Foist, who spoke for the full-time J.D. students; and Lance Howard, who spoke for the part-time division.

Isra Haider and Cynthia Adams

The law school awarded 235 Juris Doctor degrees, 12 Master of Jurisprudence degrees, and 14 Master of Law degrees. The law school also awarded 118 graduate certificates in the areas of civil and human rights law, criminal law, health law, corporate and commercial law, environmental and natural resources law, intellectual property law, and international and comparative law. In addition to their academic pursuits, the class of 2024 contributed 5, 544.68 hours of pro bono service to the central Indiana community while attending law school.

Professor Frank Sullivan and Associate Dean of Graduate and International Programs Tom Wilson served as hooding faculty for the Master of Laws and Master of Jurisprudence students. Professor Sullivan also served as hooding faculty for the Juris Doctor students, along with Professor Florence Wagman Roisman and Professor Lahny Silva. Faculty marshals were Professor James Dimitri and Professor Julie Gilgoff. Alumni marshals were Judge Fredric D. Roberson, J.D. ’02; and Anne M. Young, M.J. ’18.

« Current Headlines

IU Robert H. McKinney School of Law resources and social media channels

Give Now

Support IU McKinney Give online today!

Contact, Address and Additional Links

Planning and scheduling.

  • Course Descriptions
  • Room Calendar

Course Research/Study

  • CALI Lessons

Find McKinney

  • Law School Intranet
  • Technology/Web Request Form
  • AV Request Form (public link)

Indiana University Robert H. McKinney School of Law

Phone : 317-274-8523

  • McKinney Works Online Career Center

Texas

IC² Institute, Dell Medical School Seeking Research Proposals to Advance the Use of AI in Health Care

research proposal school dropout

The IC² Institute, in collaboration with Dell Medical School, announces a research funding opportunity aimed at UT Austin faculty and researchers on the theme of “Designing Artificial Intelligence (AI) to Mitigate Health Disparities.”

Through this 2024 call for proposals, the IC² Institute and Dell Med are offering one-year awards up to $75,000 for selected projects with a single principal investigator (PI) or up to $150,000 for collaborative projects including two or more PIs. Collaborative, multi-PI projects must include a Dell Med researcher. The deadline for concept paper submission is May 24, 2024.

Multi-disciplinary research in AI offers the potential to generate technologies producing impactful solutions to diverse and pressing problems. IC² has partnered with Dell Med to grow UT’s expertise and impact in the rapidly evolving area of Health AI — a domain that involves the development and deployment of intelligent systems capable of processing vast amounts of health data, offering diagnostic support, identifying patterns in individual or population health and assisting healthcare professionals in decision-making.

Since adopting its current mission of “innovating well-being” in 2022, the IC² Institute has been actively engaging in research and stakeholder engagement efforts that seek to improve the delivery of health care.  Executive Director S. Craig Watkins explains that engaging UT research experts on this issue holds great promise: “UT has such a strong commitment to cutting-edge research and clinical innovation. By teaming up with Dell Med to stimulate new research from our talented faculty, we can really push our knowledge about the equitable and effective use of AI in health care.” Dell Med Dean Claudia Lucchinetti, M.D., commented further:

“Innovation in technology and AI holds the potential to transform treatment and the delivery of health care.  However, ensuring that their development and application are designed with active input from stakeholders and protecting against bias are key to realizing that potential. By using a collaborative approach to the research in this space, UT is not only shaping the future of care but also making sure that patients and their families realize benefit from this innovation.”

Despite the considerable attention devoted to Health AI, some populations are overlooked in its development and potential benefits. For example, research suggests that the implicit biases in health-based data sets often influence the design of algorithmic models that can lead to disparate and negative impacts on underserved populations. Similarly, AI-informed medical devices may perform poorly when used to treat patients from under-tested populations.

To address these sorts of challenges, the IC² Institute and Dell Med seek research that falls into two broad categories:

  • Relating to the use of Health AI in safety-net contexts, research that seeks to deploy AI to better understand disparate health outcomes with greater computational nuance and precision
  • Believing that the design of future Health AI will benefit from diverse voices and expertise, research that demonstrates innovative and effective ways to engage community stakeholders in the design and deployment of Health AI.

The IC² Institute and Dell Med are open to research projects that explore health and well-being disparities in relation to geography (i.e., rural or urban); gender/sexuality; race and ethnicity; socio-economic status; and chronic diseases.

As the first part of a two-step review process, the deadline for concept paper submission is May 24, 2024. Select applicants will then be invited to participate in an interactive workshop in June and asked to submit final proposals in July. Additional details and application instructions can be found here .

Learn more about UT’s “Year of AI” initiative and other AI news and campus events at yearofai.utexas.edu .

Tags: TexasAI, Health & Well Being, IC2 Institute, Dell Medical School, Research, Technology

April 2, 2024

IMAGES

  1. Educational research project on school dropouts.docx

    research proposal school dropout

  2. Proposal Preventing School Dropout Free Essay Example

    research proposal school dropout

  3. (PDF) Teachers’ Perceptions of High School Dropout and Their Role in

    research proposal school dropout

  4. (PDF) School Dropouts

    research proposal school dropout

  5. (PDF) Factors affecting student dropout in MOOCs: a cause and effect

    research proposal school dropout

  6. Solved Can you help develop chapter III of a proposal

    research proposal school dropout

VIDEO

  1. Research Proposal (School Purposes) Group 2

  2. Adam's Surprise Proposal to Amanda at School

  3. Holt Schools Bond Proposal: Enhancing Safety and Infrastructure Without Tax Increase

  4. Why AI Is Causing College Students to Drop Out

  5. Teacher Ka Birthday

  6. Lunch Chori Ho Gaya

COMMENTS

  1. (PDF) School dropout

    show that 3.4% of students dropped out that year. Alternatively, status dropout is the percentage. of 16- to 24-year-olds who are not enrolled in school and have not earned a high school ...

  2. Understanding Why Students Drop Out of High School, According to Their

    Research on school dropout extends from early 20th-century pioneers until now, marking trends of causes and prevention. However, specific dropout causes reported by students from several nationally representative studies have never been examined together, which, if done, could lead to a better understanding of the dropout problem. ...

  3. Student Engagement and School Dropout: Theories, Evidence ...

    Research suggests that many youth who drop out of school had gone through a short- or ... three important gaps emerge between the state of empirical evidence and theoretical proposals. They pertain to (1) the trajectories of student engagement associated with student dropout, (2) the short- and long-term processes of student disengagement ...

  4. PDF Why Students Drop Out of School: A Review of 25 Years of Research

    California Dropout Research Project Report #15 October 2008 By Russell W. Rumberger and Sun Ah Lim University of California, Santa Barbara ... persistence.1 Most models have attempted to explain why students drop out of high school. Some have attempted to explain engagement, an important precursor to dropping out. Another

  5. Perspectives on the Factors Affecting Students' Dropout Rate During

    Eckstrom et al. (1986) have given a much broader perspective regarding dropout phenomena which helps us to visualize beyond the factors associated with students like the socio-cultural background of the family, the behavior of teachers and role cum distance from school, and most effectively, the persistent affordability of the family which ...

  6. Risk Factors for School Absenteeism and Dropout: A Meta-Analytic Review

    According to the National Center for Education Statistics ( 2018 ), 13% of the 8 th graders, 14% of the 10 th graders, and 15% of the 12 th graders were absent at least three days a month, and 6, 5, and 6% were absent at least five days a month, respectively. Until now, many studies on risk factors for school absenteeism and dropout have been ...

  7. Exploring statistical approaches for predicting student dropout in

    Student dropout is non-attendance from school or college for an extended period for no apparent cause. Tending to this issue necessitates a careful comprehension of the basic issues as well as an appropriate intervention strategy. Statistical approaches have acquired much importance in recent years in resolving the issue of student dropout. This is due to the fact that statistical techniques ...

  8. PDF School Dropout Indicators, Trends, and Interventions for School Counselors

    Abstract. School counselors are expected to develop programs that promote academic success. for all students, including those at risk for dropping out of school. Knowledge of key. indicators of potential dropouts and current trends in dropout prevention research may. assist school counselors in better understanding this complex issue.

  9. School dropout prediction and feature importance exploration ...

    Designing early warning systems through machine learning (ML) models to identify students at risk of dropout can improve targeting mechanisms and lead to efficient social policy interventions in education. School dropout is a culmination of various factors that drive children to leave school, and timely policy responses are most needed to address these underlying factors and improve school ...

  10. PDF Scalable Early Warning Systems for School Dropout Prevention

    country where nearly 40 percent of sixth graders drop out before completing ninth grade. The pilot program, which was implemented in 17 percent of Guatemala's primary schools and largely leveraging existing government resources, reduced the dropout rate in the transition from primary to lower secondary school by 4 percent (1.3 percentage

  11. PDF School dropout: causes, effects, prevention and intervention strategies

    Here are a few of them which we consider relevant. Finn (1989) describes two behavioural models which lead to abandoning school. The first model, frustration-self-esteem model, refers to the students who have accumulated a negative experience in school, consisting of repeated failures, low grades and low self-esteem.

  12. A critical review of the literature on school dropout

    Discussion. 7.1. Alternative credentials as an answer to school dropout. This literature review has made clear that the role of the economy, politics, and society in general is often left out of the picture. Moreover, school systems' organization and its effect on early school leaving is also still underexplored.

  13. PDF A study of the causes of dropout in public primary schools in Ndoleli

    school environmental factors affected primary school dropout rate, (2) the influence of gender on primary school dropout, (3) the effect of cultural related factors on school dropout and (4) the effects of parents' socio-economic status on school dropout in Ndoleli division, Igembe North district.

  14. PDF School Dropout Prevention

    NCES reports that on average, 3.4 percent of students who were enrolled in public or private high schools in October 2008 left school before October 2009 without completing a high school program. Broken down by race, the estimated event dropout rates were 2.4% for Whites, 4.8% for African Americans, and 5.8% for Latinos.

  15. PDF Factors Contributing to Secondary School Dropouts and the Dro- pouts

    The purpose of this study is to determine the factors that cause students drop out of school in secondary edu-cation. This is a descriptive survey study and the quantitative research method was employed. 2985 secondary school dropouts participated in the study and the data was collected through "Secondary School Dropout Qu-estionnaire".

  16. PDF INVESTIGATING THE CAUSES OF SCHOOL DROPOUT IN NINE ...

    SCHOOL DROPOUT Voice of Research | Volume 10 Issue 4, March 2022 | 154 school dropout. It was finally recommended that the Government of Rwanda and school managers should manage to put into consideration all the strategies to fight against school dropout emanated from the findings of this study. Keywords: School dropout and Nine years basic ...

  17. A methodology to design, develop, and evaluate machine ...

    School dropout is a structural problem which permanently penalizes students and society in areas such as low qualification jobs, higher poverty levels and lower life expectancy, lower pensions, and higher economic burden for governments. Given these high consequences and the surge of the problem due to COVID-19 pandemic, in this paper we propose a methodology to design, develop, and evaluate a ...

  18. School closures could mean longer commutes for students. Here's why

    In a 2017 study, researchers found that the time students spend on the school bus can mean less time for sleeping and exercise. "The research suggests that more local schools are good for students' health, and probably for their well-being," said Michael Smart, a professor of urban planning at Rutgers University…

  19. How technology is reinventing K-12 education

    For K-12 schools, this year also marks the end of the Elementary and Secondary School Emergency Relief (ESSER) funding program, which has provided pandemic recovery funds that many districts used ...

  20. UW's SER and College of Engineering and Physical Sciences Fund Energy

    Recipients have been selected from among 41 proposals for funding following a request from the University of Wyoming's School of Energy Resources (SER) to UW employees in need of energy-related laboratory equipment -- including repairs -- or supplies to conduct energy-related research. "The volume of proposals, totaling more than 10 times ...

  21. A Graduation Message for Psychology Students

    Last year, I delivered the graduation speech for Oregon State University's School of Psychological Science, and I think the message is even more relevant today. Three lessons on purpose ...

  22. SHS Certifies Four Doctoral Candidates for Graduation in the Class of

    Dr. Xiao Ding, the recipient of the Michael R. Daley Endowed Presidential Scholarship for Doctoral Students, focused her research on the impact of intervening in ecological systems, such as parent-child relationships or school culture and environment, using solution-focused brief therapy (SFBT) and analyzing students' mental health/behavioral ...

  23. Hamas Chief Says Israeli Amendments on Ceasefire Proposal Led to Deadlock

    US News is a recognized leader in college, grad school, hospital, mutual fund, and car rankings. Track elected officials, research health conditions, and find news you can use in politics ...

  24. School of Law

    Faculty Research; Our History; Recognition & Accolades; McKinney by the Numbers. Consumer Information (ABA Required Disclosures) ... and those students were celebrated during a graduation recognition celebration hosted by the law school. The law school's graduation recognition event took place at the Murat Theater at Old National Centre on ...

  25. IC² Institute, Dell Medical School Seeking Research Proposals to

    IC² Institute, Dell Medical School Seeking Research Proposals to Advance the Use of AI in Health Care . The IC² Institute, in collaboration with Dell Medical School, announces a research funding opportunity aimed at UT Austin faculty and researchers on the theme of "Designing Artificial Intelligence (AI) to Mitigate Health Disparities." ...

  26. Predicting Student Dropout Through Machine Intelligence: a Survey

    dropout using machine learning is an important area of research. It has the potential to support educational institutions in identifying at-risk students and implementing proactive measures to prevent