Abstract
Students with externalizing or internalizing behavior problems are at increased risk of underachievement and school non-completion, often due to their lower school engagement. Two studies were undertaken to assess the unique and joint (i.e., interactive) associations between behavior problems and engagement during two developmental periods; childhood and adolescence. These studies also aim to disentangle the contribution of global (externalizing and internalizing) and specific (hyperactivity/inattention, opposition/defiance, anxiety, depression) behavior problems on the global and specific aspects of student behavioral, emotional, and cognitive engagement. Study 1 was conducted among a sample of elementary school students (n = 1036; 3rd to 6th grade; mean age = 9.94 y.o.; 47.30% female; majority native Canadians) and Study 2 was conducted in secondary school (n = 1011; 7th and 8th grade; mean age = 12.93 y.o.; 55.77% female; 60.64% from immigrant background). Results of the bifactor-CFA and path analyses from both studies indicate that global externalizing behaviors were associated with lower global and specific behavioral engagement. In Study 1, global internalizing behaviors were also associated with lower global and specific cognitive engagement, whereas specific anxiety was associated with lower global and specific emotional engagement. In Study 2, specific depressive symptoms were associated with lower global and specific emotional engagement. Together, these two studies suggest that externalizing behaviors remain risk factors for student disengagement during childhood and adolescence, but that the risk posed by internalizing behaviors increases in importance for older students.
Similar content being viewed by others
Avoid common mistakes on your manuscript.
Introduction
Equity in education is an international concern (OECD 2018). In North America, disparities in upper secondary completion rates reach up to 30% between students without and with externalizing and internalizing behaviors (MEES 2019). This source of disparity is larger than that between students according to their sex (up to 7%), SES background (up to 20%), and ethnicity (up to 20%) (UNESCO 2020). Moreover, up to one-third of students enrolled in regular schools display some level of externalizing (i.e., hyperactivity, inattention, opposition, and defiance) or internalizing (i.e., anxiety and depressive symptoms) behavior adjustment problems (Olivier et al. 2018). Student disengagement in school is among the main mechanism leading to student academic failure and dropout throughout schooling (Fredricks et al. 2016), and even more strongly for youth with internalizing or externalizing behaviors (Gut et al. 2012). However, it remains unclear which of these behavior problems are more damaging for student behavioral, emotional, and cognitive engagement, as well as if the damages they contribute to are different for boys and girls, and at different developmental periods (Al-Hendawi 2012). Thus, the current study proposes to assess the nexus between behavior problems and student engagement at two developmental periods—before and after the transition to secondary school—and to assess whether some of the relations vary as a function of students’ sex.
Student Behavior Problems: The Roles of Sex and Development
Students with behavior adjustment problems are often categorized as displaying externalizing or internalizing behaviors (e.g., Caci et al. 2015). Externalizing behaviors include an array of noticeable and disruptive behaviors for one’s surroundings (Achenbach and Edelbrock 1978). Behaviors classified as externalizing typically encompass hyperactivity, attention problems, and conduct problems such as opposition and defiance (Achenbach and Edelbrock 1978). In contrast, internalizing behaviors are not as noticeable, as they pertain to more interiorized self-directed difficulties (Achenbach and Edelbrock 1978). Behaviors classified as internalizing encompass worry, sadness, anhedonia (lack of interest and pleasure), and other negative moods and emotions. Even if externalizing and internalizing behaviors are considered to be distinct (Caci et al. 2015), some children (Willner et al. 2016) and adolescents (Reitz et al. 2005) will still go on to develop both types of problems (Papachristou and Flouri 2019).
Although externalizing and internalizing behaviors often take root early in development, these two types of behavior follow distinct developmental trends. Externalizing behaviors are usually first recognized as problematic when students enter the school system and have to confront themselves with school and classroom expectations (Campbell et al. 2014). Partly because of the associated impaired self-regulation skills, hyperactive and inattentive behaviors are incompatible with expectations to pay attention to teachers’ explanations, work quietly, and interact in a prosocial manner with their peers (Campbell et al. 2014). Similar difficulties have also been observed in children with oppositional and defiant behaviors (Bierman and Sasser 2014). Such externalizing behaviors tend to persist and remain stable over time, from childhood into adolescence, especially for children who first present externalizing symptoms early in development (Bierman and Sasser 2014). Although some students develop defiant and oppositional behaviors later in adolescence, most adolescents who display them already presented some externalizing behaviors in childhood (Bierman and Sasser 2014). Moreover, boys are more numerous than girls to display externalizing behaviors, and this sex-related disparity in prevalence remains similar across development (opposition/defiance: Bierman and Sasser 2014; hyperactivity/inattention: Campbell et al. 2014).
Developmental trends are slightly different for internalizing behaviors, especially between boys and girls. Anxious and depressive thoughts and emotions are less prevalent in childhood than in adolescence. Yet, when they emerge early in development, depressive symptoms (Garber and Rao 2014) and anxiety (Vasey et al. 2014) tend to persist over time. During childhood, boys and girls present a similar risk of internalizing behaviors (depressive symptoms: Garber and Rao 2014; anxiety: Vasey et al. 2014). However, with the arrival of puberty and adolescence, rates of internalizing behaviors, especially depressive symptoms, increase among boys and girls (Nivard et al. 2017). This increase is particularly marked for girls, resulting in higher levels of anxiety and depressive symptoms among adolescent girls than boys (Garber and Rao 2014). Such developmental trends and sex differences in terms of internalizing and externalizing behaviors require further investigations to understand how behavior problems may lead students to experience negative school-related outcomes, such as lower levels of engagement (Eccles and Roeser 2009).
Finally, co-occurring externalizing and internalizing problems may be more problematic when problems start early in the development (Nivard et al. 2017). Such co-occurring problems also tend to be more stable over youth development (Willner et al. 2016). However, it is unclear whether co-occurring internalizing and externalizing behaviors are more salient in boys or girls, as studies on this subject have yielded mixed results. Indeed Olivier et al. (2018) found that boys were more numerous to display externalizing behaviors, whereas girls were more numerous to display internalizing behaviors. Papachristou and Flouri (2019) found that boys were more likely to present higher levels of externalizing and internalizing behaviors. Finally, Reitz et al.’s (2005) results show that girls have higher levels of internalizing behaviors but that boys and girls do not differ in terms of externalizing behaviors. Despite the general recognition that externalizing and internalizing behaviors are independent risk factors, it is generally accepted that an accumulation of such problems is likely to bring about additional risks (Evans et al. 2013). As youth spend an important proportion of their time ins school, these risks are likely to threaten their adjustment to school, particularly their engagement.
Student Engagement: Components and Developmental Trends
Student engagement reflects students’ underlying motivation for their schoolwork (Skinner et al. 2009). As such, engagement is considered to be “motivation in action” (Skinner et al. 2009). According to the widely accepted three-dimensional definition proposed by Fredricks et al. (2004; see also Fredricks et al. 2016), student engagement encompasses behavioral, emotional, and cognitive components, all of which are necessary for active student investment in their schoolwork. Behavioral engagement refers to students’ conduct and actions, including active participation, attendance, as well as compliance with classroom rules and teacher instructions (Fredricks et al. 2004). Emotional engagement encompasses students’ affective reactions to the learning process and classroom environment, including their interest, enjoyment, and happiness (Fredricks et al. 2004). Cognitive engagement includes students’ thought processes that promote their dedication of efforts to learning and mastery of school-related tasks and content. As such, cognitive engagement encompasses self-regulated and deep-processing strategies, such as reformulating ideas in one’s own words, identifying important information, and trying different strategies to solve problems (Fredricks et al. 2004). Student engagement seems to remain stable across development for a majority of students, both in children (Archambault and Dupéré 2016) and adolescents (Wang and Eccles 2012). In addition, girls tend to display slightly higher engagement levels than boys, regardless of the developmental period considered, particularly for language classes (Archambault et al. 2009). Apart from these different levels of engagement, boys and girls seem to follow similar developmental trends (Li and Lerner 2011).
Behavior Problems and Student Engagement: Stage-Environment Fit Theory
According to Stage-Environment Fit Theory (Wigfield and Eccles 2000), students display an optimal engagement when the school environment (the school policies, pedagogical practices, climate, etc.) offers a level of support that matches their developmental needs for belongingness, competence, autonomy, and security at each developmental stage. For most students, the fit between their school settings and their own needs as developing individuals is satisfying enough for them to maintain adequate engagement levels, leading to academic success. Students suffering from externalizing and internalizing problems usually need intensified support from their schools to display a level of engagement comparable to that of their peers. Yet, the support required for these youth is not always within reach of what schools can offer. Besides, despite this greater need, these students might end up receiving less support than their peers. Some of their behaviors can interfere with efficient classroom-functioning and their own individual functioning (Campbell et al. 2014). As such, students with externalizing and internalizing problems are likely to experience a greater level of stage-environment mismatch at school, possibly leading to lower school engagement levels.
Given that school resources are limited identifying which types of adjustment problems are the most important risk factors for students’ school engagement might help better targeting students with the most pressing needs at different developmental periods. Moreover, these risks are likely to change at different stages of students’ schooling, and possibly between boys and girls. The next section reviews studies assessing the nexus between adjustment problems and student engagement by contrasting the contribution of externalizing and internalizing behaviors, during childhood versus adolescence, and between boys and girls.
Behavior Problems and Student Engagement: An Empirical Review
Table 1Footnote 1 reviews previous studies that have focused on the school engagement of students displaying varying levels of externalizing and internalizing behaviors. A general overview of this table first depicts that studies assessing externalizing behaviors mainly focus on childhood, whereas studies assessing internalizing behaviors mainly focus on adolescence. When assessed separately from internalizing behaviors, externalizing behaviors have been systematically found to be associated with lower levels of behavioral engagement in children (see Table 1: Archambault et al. 2017; Demaray and Jenkins 2011; Junod et al. 2006; Moilanen et al. 2010; Olivier and Archambault 2017; Plamondon and Martinussen 2019; Volpe et al. 2006) and adolescents (Tian et al. 2017; Wang and Fredricks 2014). Despite these consistent findings, the relations between externalizing behaviors and the emotional and cognitive components of student engagement are far less clear. For instance, Archambault et al. (2017) found that students with high levels of opposition/defiance reported low emotional engagement, but Olivier and Archambault (2017) found no such association for students displaying hyperactivity/inattention. Among secondary school students, studies have considered emotional and cognitive engagements in a global measure of student engagement and found that hyperactive-inattentive and opposition/defiance behaviors led to a lower global engagement (see Table 1: Nguyen et al. 2019; Tian et al. 2017; Wang and Fredricks 2014). Finally, the few studies that have assessed possible sex differences found that externalizing behaviors impacted boys’ and girls’ engagement similarly (see Table 1: Archambault et al. 2017; Demaray and Jenkins 2011; Olivier and Archambault 2017).
Most research focusing on the association between internalizing behaviors and student engagement has been conducted among samples of adolescents and has focused primarily on the effects of depressive symptoms. Among children, Kurdi and Archambault (2020) found that anxiety does not directly lead to decreased behavioral, emotional, and cognitive engagement, neither in boys nor in girls. In adolescence, studies tend to report that depressive symptoms negatively impact students’ behavioral, emotional, or global engagement (see Table 1: Dorio et al. 2019; Fiorilli et al. 2017; Garvik et al. 2014; Wang and Peck 2013; Wang et al. 2015). Although Dorio et al. (2019) found that depressive symptoms lead to a larger decrease in global engagement for girls, Derdikman-Eiron et al. (2011) instead found that internalizing behaviors, including anxiety and depressive symptoms, led to a lower behavioral engagement for all youth, but more pronounced for boys. Thus, there is no consensus on a sex that might be more at risk. In general, Garvik et al. (2014) warn that the effect sizes of depressive symptoms remain small, suggesting that even if internalizing behaviors seem to hamper students’ engagement, most of them appear to be able to manage their symptoms to maintain adequate engagement. However, as for externalizing behaviors, associations involving the emotional and cognitive component of engagement are not as well established as those involving the behavioral dimension across developmental levels and samples of boys and girls.
Six studies have assessed the simultaneous repercussions of externalizing and internalizing behaviors on student engagement. Among children, Searle et al. (2013) found that hyperactivity-inattention, but not opposition-defiance and depressive symptoms were associated with lower global engagement. In contrast, Baker et al. (2008) found that internalizing behaviors, but not externalizing behaviors, led to lower behavioral engagement. Curhan et al. (2020) found similar results for adolescents’ emotional engagement. Li et al. (2008) and Li and Lerner (2011), more specifically, studied opposition/defiance and depressive symptoms among adolescents. They found that these two adjustment problems led to decreases in behavioral and emotional engagement. Finally, Olivier et al. (2018, 2020) found that an externalizing profile of problems led to lower behavioral, emotional, and cognitive engagement for girls but not for boys. Girls with an internalizing profile of problems also reported a lower engagement. Yet, boys, but not girls, were likely to display externalizing and internalizing problems simultaneously, which led them to report a lower behavioral and emotional engagement. This last result suggests that there may be negative repercussions of accumulating behavior problems. For instance, Eisenberg et al. (2009) and Wang et al. (2016a, 2016b) assessed students’ self-regulatory abilities, a component of cognitive engagement. They found that youth displaying externalizing behaviors, or a combination of externalizing and internalizing behaviors, were more at risk of showing lower levels of such abilities than youth reporting only internalizing behaviors, or no problem. Such results call for a more in-depth assessment of the combined, and possibly harmful, influence of externalizing and internalizing behaviors on student engagement.
Finally, although the studies reported in Table 1 have alternatively relied on subject-specific (i.e., math or language) measures of engagement, this methodological consideration does not seem to have impacted the results. This observation could be partly explained by all of these studies having relied on domain-general (rather than subject-specific) measures of behavioral problems. In addition, although student engagement levels may differ across subjects, subject-specific engagement seems to remain importantly driven by motivation, affects, and engagement encompassing several school subjects (e.g., Gogol et al. 2017).
Toward a Better Understanding of the Global and Specific Components of Externalizing Behaviors, Internalizing Behaviors, and School Engagement
A precise assessment of the unique and combined effects of distinct behavior problems on student’s engagement involves accounting for the high rate of co-occurrence between ratings of hyperactivity/inattention and opposition/defiance (Campbell et al. 2014), as well as between ratings of anxiety and depressive symptoms (Vasey et al. 2014). This high level of co-occurrence makes it difficult to precisely account for the role of global levels of externalizing/internalizing behaviors, relative to the role of specific behavior problems. A typical approach would be to focus on global levels of externalizing/internalizing behaviors, thus neglecting the possible role of specific types of behavioral problems. An alternative approach could focus on specific behavioral problems, which then carries the risk of obtaining results tainted by the role of the unmodelled global levels of externalizing/internalizing behaviors. Indeed, multivariate analyses estimate the role played by each predictor net of what it shares with the others. When these analyses involve highly related predictors (e.g., anxiety and depressive symptoms), this shared component tends to be quite large. Partialling it out often leads to an overestimation of the role played by one component, and to an underestimation of the role played by the other components. A similar issue occurs when considering students’ ratings of behavioral, emotional, and cognitive engagements, which also tends to be highly correlated with one another, and yet to each have their own unique character (Wang et al. 2019).
A bifactor approach makes it possible to identify a global factor (G-factor) together with subscale-specific orthogonal factors (S-factors). The G-factor reflects the commonality shared among all items (e.g., global levels of internalizing behaviors across dimensions). The S-factors are orthogonal (e.g., specific levels anxiety and depressive symptoms) and reflect the variance shared across items forming each subscale beyond that already explained by the G-factor (Morin et al. 2016). In plain language, this approach separates ratings into independent (i.e., uncorrelated) components. Thus, this approach simultaneously considers the role of students’ global levels of internalizing or externalizing behaviors, together with the unique predictive value of anxiety, depressive symptoms, hyperactivity/inattention, and opposition/defiance beyond these global levels. The ability of this approach to achieve a more accurate representation of behavior problems (e.g., Caci et al. 2015; Caspi et al. 2014) and engagement (e.g., Wang et al. 2016a, 2016b, 2019) has been previously demonstrated. However, this approach has yet to be implemented in research focusing on the nexus between behavior problems and engagement (see Table 1).
Current Study
Scholars and school practitioners all agree that students’ externalizing and internalizing behavior problems cause a threat to their active engagement in school. However, it remains unclear if one type of behavior problem is a more important risk factor for student engagement and how these risks may differ between boys and girls and between different developmental periods. This research assesses the relations between students’ externalizing (global, hyperactivity/inattention, and opposition/defiance) and internalizing (global, anxiety, and depressive symptoms) behaviors and their global and specific levels of engagement (behavioral, emotional, and cognitive). This study also investigates whether these associations are similar or differ across samples of elementary (Study 1; grades 3 to 6) and secondary (Study 2; grades 7 and 8) students, as well as between boys and girls. Studying these associations before and after the transition to secondary school can inform about student adjustment during this period, where they are known to be particularly at risk of disengagement (Eccles and Roeser 2009). More specifically, the present research pursues two objectives, via two independent studies:
The first objective of this study is to investigate the associations between global and specific levels of externalizing and internalizing behaviors and student engagement. This study anticipates that both types of behavioral problems will share negative associations with student engagement levels, both in elementary and secondary school students. Yet, existing results are mixed regarding which of these problems has the strongest association with which dimension of engagement. As such, the differential effects of global and specific levels of internalizing and externalizing behaviors in the prediction of global relative to specific components of student engagement is left as an open research question. Moreover, although different developmental trends have been reported between boys and girls, most studies that have assessed sex differences in the nexus linking behavior problems to student engagement have found no such differences (see Table 1). Thus, the study anticipates that these associations will also be similar for boys and girls.
The second objective of this study is to verify whether there are interactive effects between behavior problems in predicting student engagement. Finding a significant interaction would be consistent with the presence of a risk accumulation effect due to the co-occurrence of various behavior problems. More specifically, the study assesses the interactions between high global levels of externalizing and internalizing behaviors; between high global levels of externalizing behaviors and high specific levels of internalizing behaviors (anxiety, and depressive symptoms); between high global levels of internalizing behaviors and high specific levels of externalizing behaviors (hyperactivity/inattention, and opposition/defiance); and between high specific levels of externalizing and internalizing behaviors. Although the co-occurrence of behavior problems is reported in several studies, it is unclear if this risk accumulation causes extra threats to student engagement. Still, co-occurring externalizing and internalizing behaviors seem more problematic when they appear early in development. Thus, the study hypothesizes that this risk accumulation (i.e., combining multiple types of behavioral problems) might be associated with a further reduction in engagement, particularly in the elementary school sample.
Study 1 (Elementary School Sample)
Methods
Sample and procedure
Study 1 relies on a sample of 1036 3rd to 6th grade students recruited in seven elementary schools from one school board located in the Canadian province of Quebec. The majority of students were Caucasian and came from middle-class families, which is representative of the student population outside of the Montreal area (MEES 2019). Students were, on average, 9.94 years old (SD = 1.30), and 47.30% of them were girls. This project was approved by the University’s research ethics committee. Prior to data collection, the research team obtained active parental, student, and teacher consent. In November 2011, students answered a 45 min computerized questionnaire on their school experiences, including all measures used in the present study. During data collection, each classroom was supervised by two trained research assistants, and teachers used this time to complete a paper questionnaire on the behaviors and performance of each of their studentsFootnote 2.
Measures
Externalizing behaviors
Teachers rated each student using two scales from the French adaptation (Capron et al. 2007) of the Strengths and Difficulties Questionnaire (Goodman 2001): Hyperactivity/Inattention (five items; α = 0.764; e.g., “This child is restless, overactive, cannot stay still for long”) and Oppositional/Defiant (four itemsFootnote 3; α = 0.725; e.g., “This child often loses temper”). Each item was rated on a three-point scale (1—not true to 3—certainly true).
Internalizing behaviors
Student rated their own levels of internalizing behaviors using two scales from the French adaptation (Trembaly et al. 1987) of the Preschool Behavior Questionnaire (Hoge et al. 1985): Anxiety (four items; α = 0.743; e.g., “You worry that you are not as good as other students”) and Depressive symptoms (three items; α = 0.779; e.g., “You are unhappy or sad”). Each item was rated on a three-point response scale (1—not true to 3—certainly true).
Student engagement
Students completed the Dimensions of School Engagement Scale (Archambault and Vandenbossche-Makombo 2014) in relation to their language class (i.e., French). The behavioral (α = 0.685; e.g., “I follow my teachers’ instructions”), emotional (α = 0.770; e.g., “I think that reading and writing assignments are interesting”), and cognitive (α = 0.737; e.g., “When I finish an assignment, I check to make sure that I did not make mistakes”) engagement subscales each comprised three items rated on a five-point scale (1—not at all to 5—very much).
Covariates
Students self-reported their sex (0 = male; 1 = female), grade level (3rd, 4th, 5th, or 6th grade), and family status (0 = nuclear family; 1 = other). Teachers reported the achievement of their students in language class (i.e., French) by comparing each of them to the class average on a scale ranging from (1) significantly below average to (5) significantly above average (Duncan et al. 2007).
Analyses
All analyses were performed with Mplus 8.2 (Muthén and Muthén 2019). Alternative measurement models were first estimated and contrasted to establish the optimal factor solution (CFA or bifactor-CFA) for each set of variables (externalizing behaviors, internalizing behaviors, and student engagement) (see Appendix 1 for details). The selection of the optimal and most invariant measurement model for all constructs was followed by the estimation of predictive models designed to assess the role of externalizing and internalizing behaviors in the prediction of student engagement. Given the complexity of the measurement models estimated in this study, it was not possible to estimate these predictive models using a fully latent approach. Rather, the study relied on factor scores saved from the optimal measurement model identified previously. This allowed relying on the Maximum Likelihood Robust (MLR) estimator for these models, together with Full Information Maximum Likelihood procedures (Enders 2010) to handle missing data (5.77 to 28.10%) on the covariates.
A first predictive model including only the covariates (sex, grade, family status, and achievement) and their association with student engagement was then estimated. To maximize model parsimony, only the covariates associated in a statistically significant manner with each outcome were retained in the main analyses in an outcome-specific manner. In a second model, externalizing and internalizing behaviors were added as additional predictors. In a third model, two-way interactions between the externalizing and internalizing behaviors factors were also included in order to address Objective 2. Statistically significant interaction effects were interpreted by the examination of simple slopes depicting the effect of the predictor at different levels (−1 SD, M, and +1 SD) of the moderator (Marsh et al. 2013). Finally, the predictive equivalence of these relations across sexes was assessed through the following sequence of models: (i) configural equivalence (same predictive model with no additional constraint); (ii) equivalence of the regression slopes; (ii) equivalence of the regression intercepts; (iii) equivalence of the regression residuals.
Model fit was assessed using the chi-square statistic (χ2), the Root Mean Square Error of Approximation (RMSEA), the Comparative Fit Index (CFI), and the Tucker-Lewis Index (TLI) (Marsh et al. 2005). RMSEA values smaller than 0.08 and 0.06 respectively suggest acceptable and excellent model fit. Values above 0.90 and 0.95 for the CFI and TLI respectively indicate adequate and excellent model fit. In tests of measurement invariance and predictive equivalence, increases in RMSEA of more than 0.015 and decreases in CFI and TLI of more than 0.010 were considered to indicate non-invariance (Chen 2007).
Results
Correlations between the various factor scores and covariates are reported in Table 2. Results from the model including only covariates are reported on the top section of Table 3. Based on this model, family structure was removed from further analyses, and only significant paths involving student sex, grade level, and achievement were kept in the following models (sex and grade level in the prediction of global engagement and specific emotional engagement, and achievement in the prediction of global engagement, and specific behavioral and emotional engagement). The main predictive model had an excellent level of fit to the data (χ2 = 1.453, df = 5, p = 0.918; RMSEA = 0.000; CFI = 1.000; TLI = 1.042). These results (see Table 3) indicate that the Externalizing behaviors G-factor was negatively associated with the Engagement G-factor and Behavioral engagement S-factor. The Internalizing behaviors G-factor was also negatively associated with the Engagement G-factor and Cognitive engagement S-factor. Moreover, the Anxiety S-factor was associated with lower levels on the Engagement G-factor and the Emotional engagement S-factor. The Hyperactivity/Inattention, Oppositional/Defiant, and Depressive symptoms S-factors did not contribute to the prediction of any engagement factor. Adding interactions between G-factors, between G-factors and S-factors, and between S-factors did not further contribute to explaining any of the engagement G- and S-factors. Tests of predictive equivalence across sexes supported the equivalence of the regression slopes, but suggest sex differences related to the regression intercepts associated with the Engagement G-factor and the Emotional engagement S-factor. More precisely, these regression intercepts were respectively 0.449 and 0.466 SD higher among girls relative to boys, suggesting that girls tend to present higher levels (roughly 0.5 SD) on these S-factors relative to boys presenting matching levels of internalizing and externalizing behaviors. Likewise, the results also revealed sex differences related to the regression residual of the Cognitive engagement S-factor, showing that the model was able to explain 2.2% of the variance of Cognitive engagement for boys compared to 3.0% for girls.
Study 2 (Secondary School Sample)
Methods
Sample and procedure
Study 2 relies on a sample of 1011 7th and 8th grade secondary school students recruited in four schools located in the Canadian province of Quebec. Although collected in the same Canadian province, this sample is independent from, and unrelated to, the sample used in Study 1. Students were on average 12.93 years old (s.d. = 0.76), 55.77% of them were girls, 26.50% were first-generation immigrants (born abroad), and 34.14% were second-generation immigrants (born in Canada, with one or both parents born abroad), and 39.26% were third-generation plus. This project was approved by the University’s research ethics committee. Active written parental consent was obtained for all students. Students also gave their active consent to participate. In Fall 2018, students answered a 45 min computerized questionnaire on their well-being and school experiences including all measures used in the present study. During data collection, each classroom was supervised by a trained research assistantFootnote 4.
Measures
Externalizing behaviors
Student self-reported their externalizing behaviors using two scales from the French adaptation (Capron et al. 2007) of the Strengths and Difficulties Questionnaire (Goodman 2001): Hyperactivity/Inattention (five items; α = 0.654; e.g., “I am restless, I cannot stay still for long”) and Oppositional/Defiant (five items; α = 0.582; e.g., “I lose my temper easily or often get angry”). Each item was rated on a three-point scale ranging from (1) not true to (3) certainly true.
Internalizing behaviors
Student self-reported their internalizing behaviors using two scales. Anxiety was assessed using nine items (α = 0.864; e.g., “You worry that you are not as good as other students”) from the Screen for Child Anxiety Related Emotional Disorders (Birmaher et al. 1999) validated in French (Blais et al. 2002). These items were rated using a three-point scale ranging from (1) almost never to (3) often. Depressive symptoms were measured using six items (α = 0.932; e.g., “You feel sad”) from a short French version (Blais et al. 2002) of the Center for Epidemiologic Studies Depression scale (Radloff 1977). These items were rated on a six-point scale ranging from (1) never to (6) always.
Student engagement
Students completed a measure of engagement directly developed and validated in French (Archambault et al. 2009). This scale is not specific to any school subject and covers three dimensions: Behavioral (four items; α = 0.712; e.g., “I have disrupted the classroom on purpose”), emotional (six items; α = 0.851; e.g., “I enjoy what we do in school”), and cognitive (seven items; α = 0.895; e.g., “I take time to make sure that I understand assignments”). Items were rated on a five-point response scale ranging from (1) strongly disagree to (5) strongly agree.
Covariates
Students self-reported their sex (0 = male; 1 = female), grade level (7th or 8th grade), family status (0 = nuclear family; 1 = other), grade retention (0 = never; 1 = repeated a grade at least once), and immigration status (0 = non-immigrant; 1 = 1st and 2nd generation immigrants).
Analyses
The same procedure as in Study 1 was applied to Study 2. Given that Study 2 was conducted among a multiethnic sample, measurement invariance and predictive equivalence were also tested between groups of immigrant (1st and 2nd generation) and non-immigrant students (3rd generation+). All measures were student-reported using a computerized method allowing for no missing data.
Results
Correlations between factors and covariates are reported in Table 2. Results from the model including only covariates are reported in the bottom section of Table 3. Based on this model, only significant paths involving student sex, grade level, retention, family structure, and immigration status were kept in the following models (sex in the prediction of all outcomes, grade level in the prediction of global engagement and specific behavioral and emotional engagement, family status in the prediction of global engagement and specific emotional engagement, retention in the prediction of specific behavioral engagement, and immigration status in the prediction of all three specific engagement dimensions). The predictive model had an excellent level of fit to the data (χ2 = 9.889, df = 7, p = 0.195; RMSEA = 0.020; CFI = 0.998; TLI = 0.984). These results (see Table 3) indicate that the Externalizing behaviors G-factor and Hyperactivity/Inattention S-factor were negatively associated with the Engagement G-factor and with the Behavioral engagement S-factor. The Oppositional/Defiant S-factor was also negatively associated with the Behavioral engagement S-factor. The Internalizing behaviors G-factor and the Anxiety S-factor were not directly associated with any dimension of engagement. The Depressive symptoms S-factor negatively contributed to the Engagement G-factor and to the Emotional engagement S-factor. Finally, no variable contributed to the prediction of the Cognitive Engagement S-factor.
When adding interactions between G-factors, between G-factors and S-factors, and between S-factors, model fit remained excellent (χ2 = 10.460, df = 7, p = 0.164; RMSEA = 0.022; CFI = 0.997; TLI = 0.972) and the significance of all direct effects remained unchanged. These results revealed two statistically significant interactions. Thus, the interaction between the Internalizing behaviors G-factor and the Hyperactivity/Inattention S-factor was significantly associated with the Engagement G-factor (β = 0.086, p = 0.013) and with the Emotional engagement S-factor (β = 0.091, p = 0.009). The results from simple slopes analyses are graphically presented in Figs. 1 and 2. These results indicate that, for students with a low level on the Hyperactivity/Inattention S-factor, the Internalizing behaviors G-factor was associated with a lower level on the Engagement G-factor and the Emotional engagement S-factor. For students with a high level on the Hyperactivity/Inattention S-factor, the Internalizing behaviors G-factor was associated with a higher level on these two dimensions.
Results from the global internalizing behaviors (G-factor) × specific hyperactivity/inattention interaction (S-factor) in the prediction of global engagement levels (G-factor) in the secondary school sample
Results from the global internalizing behaviors (G-factor) × specific hyperactivity/inattention interaction (S-factor) in the prediction of specific emotional engagement levels (S-factor) in the secondary school sample
Tests of predictive equivalence across immigration status supported the equivalence of the regression slopes and regression residuals, but not of the regression intercepts. More precisely, immigrants tended to present lower intercepts on the Behavioral engagement S-factor (−0.233 SD) relative to non-immigrants presenting similar levels of internalizing and externalizing behaviors. Tests of predictive equivalence also revealed sex differences related to some regression slopes and intercepts but supported the equivalence of the regression residuals. In fact, a single regression slope, related to the association between immigration status (a covariate) and the Emotional engagement S-factor was found to differ as a function of sex, being statistically significant for girls (b = –0.197, s.e. = 0.057, p < 0.001, β = –0.127; showing that non-immigrant girls had higher specific levels of emotional engagement than immigrant girls), but not for boys (b = 0.023, s.e. = 0.060, p = 0.701, β = 0.014). In terms of regression intercepts, girls presented higher scores on the Behavioral (0.140 SD) and Cognitive (0.125 SD) engagement S-factor relative to boys presenting similar levels of behavioral problems.
Discussion
Students with externalizing or internalizing behavior problems are at increased risk of not completing upper secondary school (MEES 2019), partly due to their lower engagement (Gut et al. 2012). Considering that these students represent an important proportion of those enrolled in regular schools (Olivier et al. 2018), researchers and practitioners strive to identify priority prevention and intervention targets to minimize their risk of disengagement. Behavior problems can be conceptualized as global (externalizing or internalizing) and specific (hyperactivity/inattention, opposition/defiance, anxiety, and depression) difficulties, which are likely to play a distinct role on student disengagement (Caci et al. 2015). However, it is still unclear which type of behavior problem, internalizing or externalizing, are associated with higher risk of disengagement for boys and girls at different developmental periods. Whether and how the accumulation of different behavior problems could pose an additional risk for student school functioning also remain unknown (Eisenberg et al. 2009). To identify more specific intervention targets, this study first aimed to disentangle the global and specific contribution of externalizing and internalizing behavior problems on the behavioral, emotional, and cognitive engagement of elementary and secondary school students. Second, the study assessed the association between global and specific behavior problems with student engagement, while also contrasting the results between boys and girls.
More specifically, this study sought to investigate the combined role played by global and specific levels of externalizing and internalizing behaviors in the prediction of students’ global and specific levels of school engagement during the end of elementary school (Study 1) and the beginning of secondary school (Study 2). The study also considers the possible effects of sex (Study 1 and 2) and immigration status (Study 2 only). The results show that the global aspects of student externalizing and internalizing problems are as important as their specific facets to understand threats to student engagement. Although not longitudinal in nature, the juxtaposition of results obtained from these two studies suggests that developmental trends might influence the nature and strength of the associations between students’ behavior adjustment problems and their levels of school engagement.
Externalizing and internalizing behaviors seem to function slightly differently according to the developmental period. First, externalizing behaviors appeared important risk factors for students’ global engagement levels and their specific levels of behavioral engagement across the elementary and secondary school years. These effects appeared not to be limited to students’ global levels of externalizing behaviors during the elementary school years, but also encompass their specific hyperactivity/inattention and opposition/defiance during the secondary school years. Second, the role of internalizing behaviors changed more importantly across these two developmental periods. Thus, students’ global levels of internalizing behaviors and their specific levels of anxiety were found to be associated with lower global levels of engagement during the elementary school years. During this period, students displaying high anxiety, beyond their global internalizing behaviors, also reported lower specific levels of emotional engagement. In contrast, during the secondary school years, only students displaying high specific levels of depressive symptoms, beyond their global level of internalizing behaviors, reported lower global levels of school engagement and lower specific levels of emotional engagement.
Overall, the results suggest that at least some of the global and specific facets of behavior adjustment problems appear to play a role reducing some aspect of student engagement over the course of youth development. However, the mechanisms underpinning these effects might be submitted to developmental differences. Moreover, beyond their negative association with students’ global levels of school engagement, externalizing behavior problems also shared negative associations with students’ specific levels of behavioral engagement, whereas internalizing behaviors also shared negative associations with students’ specific levels of emotional engagement. These specific associations are consistent with the view that externalizing behaviors may prevent the optimal behavioral investment of students because of self-regulation difficulties (e.g., Campbell et al. 2014). Internalizing behaviors may rather impede students’ positive emotional investment because of cognitive biases or emotional dysregulation, which may alter their cognitive availability (Garber and Rao 2014).
Externalizing Behaviors and Student Engagement
As expected, externalizing behaviors were found to be associated with lower levels of student engagement in a way that was highly similar across the elementary and secondary school samples. More precisely, global levels of externalizing behaviors were associated with lower global levels of school engagement and lower specific levels of behavioral engagement in both samples, as well as for boys and girls. Thus, irrespective of developmental stage and sex, externalizing behaviors seem to remain incompatible with the behavioral demands of a classroom, as well as with displaying a globally engaged attitude toward school.
Beyond the effects of students’ global levels of externalizing behaviors, their specific levels of hyperactivity/inattention and oppositional-defiance were found to further contribute to reducing their global levels of engagement as well as their specific levels of behavioral engagement in the secondary school sample. There are a few possible explanations for these observations. First, as students enter secondary school, they are exposed to more diverse and complex subjects but also have to navigate a new environment, while receiving less intensive support from teachers and school professionals (Eccles and Roeser 2009). In this context, high specific levels of hyperactivity/inattention and opposition/defiance, and the related self-regulation impairments, may become additional threats to their full engagement in school. From a prevention perspective, these results suggest that targeting youth displaying very high specific levels of hyperactivity/inattention or opposition/defiance, as it is often done in various school systems, may prove fruitful. However, these results also indicate that students displaying a wider range of externalizing behaviors may benefit from preventive interventions, even if their levels of behavioral problems do not appear to be as extreme as that of their peers in specific areas.
Second, the source of information used to obtain ratings of externalizing behaviors in the present study may play a role in these differential findings. In the secondary school sample, students self-reported their own externalizing behaviors, whereas in the elementary school sample, teachers rated the externalizing behaviors of their students. Teachers are known to rely on a more global or holistic assessment process when asked to describe their students’ behaviors, which might have made it harder to detect effects associated with more specific types of behaviors in this sample (Flake and Petway 2019). In contrast, secondary school students were asked to rate both their externalizing behaviors and their school engagement, which might have slightly inflated the associations between these two types of constructs.
Another noteworthy difference is related to the items used to measure student behavioral engagement. In the elementary school sample, these items mainly focused on observable deviations from complying with classroom expectations. In contrast, in the secondary school sample, these items focused on more widespread types of disruptive behaviors. This more generic coverage may make it easier to differentiate the predictive role of global levels of externalizing behaviors from that of behaviors more specifically related to hyperactivity/inattention and opposition/defiance. In contrast, it might be harder to break down the effects of classroom compliance into specific dimensions of externalizing behavior. Future research is needed to assess the plausibility of these alternative explanations more systematically. However, given the consistency of results obtained across samples and sex, these results provide substantial evidence that global externalizing behaviors seem to represent threats to student engagement both before and after the transition to secondary school.
Internalizing Behaviors and Student Engagement
As expected, internalizing behaviors also appeared to threaten student engagement. However, the mechanisms underpinning these effects seem to change over the course of development. In the elementary school sample, students’ global levels of internalizing behaviors were found to be associated with lower global levels of engagement and with lower specific levels of cognitive engagement. This result indicates that the combination of anxiety and depressive symptoms seems to prevent students from being fully engaged in their learning experiences. Moreover, students’ global level of internalizing behaviors also prevented their full cognitive engagement. Although unexpected, this result is consistent with the view that students’ emotional dysregulation may lead to difficulties in self-regulated and deep processing learning strategies.
However, the size of these effects was not as large as those associated with global levels of externalizing behaviors, even though teachers’ reported students’ externalizing behaviors, whereas students self-reported their levels of internalizing behaviors and engagement in this sample. Effect size disparities were also reported in earlier studies (Searle et al. 2013), suggesting that the stronger effects of externalizing, relative to internalizing, behaviors on student engagement may not be specific to the present study. In addition, specific levels of anxiety were also found to be related to lower levels of global engagement and with lower specific levels of emotional engagement among elementary students. This suggests that in younger students, anxiety may play a specific role in impairing student engagement, especially when the emotional aspects of student engagement are considered (Kurdi and Archambault 2020). The higher prevalence of anxiety than depressive symptoms in younger children (Vasey et al. 2014) might explain why high specific levels of anxiety appear to be significant threats to global levels of engagement during this developmental period. Indeed, students might not yet have learned to appropriately self-regulate their anxious thoughts and emotions.
In the secondary school sample, only specific levels of depressive symptoms were found to be directly associated with students’ lower global levels engagement and with lower specific levels of emotional engagement. These two dimensions are the same that were found to be affected by specific levels of anxiety in the elementary school sample. Contrasting these results suggests a developmental difference, which may be related to the greater prevalence of anxiety (relative to depressive symptoms) in younger samples (Vasey et al. 2014), and to the increasing rates of depressive symptoms observed among adolescents (Garber and Rao 2014). Students displaying anxiety or a combination of anxiety and depressive symptoms (i.e., global levels of internalizing behaviors) at a younger age may have progressively learned to control and self-regulate their emotions and thoughts in order to protect their ability to stay engaged at school despite these negative emotions. This would be detected as a lack of associations between these types of internalizing behaviors and engagement in the secondary school sample. Such results are consistent with those previously reported by Garvik et al. (2014), demonstrating only small associations between internalizing behaviors and student engagement among a sample of adolescents. In contrast, students’ experiencing a rise in depressive symptoms during adolescence may not have learned to control the impact of these symptoms as their peers with an earlier onset of internalizing behaviors. This would then explain that the negative association found for anxiety in younger children shifts to depressive thoughts in young adolescents, who have yet to learn to manage these negative feelings.
Accumulation of Behavioral Risks
The results did not uncover the expected effect of an accumulation of behavioral risks in the prediction of engagement levels in any of the samples or subsamples considered in the present study. Moreover, the interaction effects found within the secondary school sample showed that for students presenting high levels of hyperactivity/inattention, global levels of internalizing behaviors were associated with higher global levels of engagement and with higher specific levels of emotional engagement. In contrast, for students displaying low levels of hyperactivity/inattention, global levels of internalizing behaviors were associated with lower global levels of engagement, and with lower specific levels of emotional engagement.
This unexpected interaction could indicate two possible phenomena. First, it is possible that these two behavioral risk factors, when accumulated, tap into incompatible processes. As such, having a tendency to be very active and distracted while simultaneously being concerned, nervous, and sad might exert counterbalancing effects, allowing students to maintain adequate engagement levels. Being very agitated could be a strategy for these students to deal with anxious and depressive thoughts. In a review of the literature, Jarrett et al. (2014) have shown that internalizing behaviors tend to lead to lower levels of externalizing behaviors (also see Morin et al. 2017). These observations thus suggest at least some level of incompatibility between externalizing and internalizing behaviors. Second, it is possible that secondary school students characterized by multiple types of behavioral problems might have been targeted for school-based interventions or provided access to extra support mechanisms to maintain a satisfactory level of engagement. A study by Olivier et al. (2018) did find that boys who presented a combination of externalizing and internalizing behaviors were less at risk of disengagement than students displaying only one type of such behaviors. In any case, future research is needed to investigate the plausibility and complementarity of these two explanatory hypotheses.
Universality of Behavior Problems as Risk Factors
The associations between students’ externalizing and internalizing behavior problems and engagement were found to be the same between boys and girls, and between students from immigrant and non-immigrant backgrounds. This suggests that, although the levels of behavior problems and engagement may vary as a function of sex or immigration status (see Appendices 2 and 3), externalizing and internalizing behavior problems are likely universal risk factors. As such, these risks should not be subject to differential treatment from school professionals. Similarly, some studies found that behavior problems influence boys and girls equally (opposition/defiance: Archambault et al. 2017; hyperactivity/inattention: Demaray and Jenkins 2011; depressive symptoms: Garvik et al. 2014; anxiety: Kurdi and Archambault 2020) as well as immigrants and non-immigrants (aggression: Konold et al. 2017; hyperactivity/inattention: Tardif-Grenier et al. 2019). Yet, these results are far from consensual as others found boys and girls (anxiety and depressive symptoms: Derdikman-Eiron et al. 2011; externalizing and internalizing: Olivier et al. 2018) and youths from different cultures (anxiety: Gillen-O’Neel et al. 2011) to be differently impacted, especially by internalizing behaviors. Globally, these results call for a more systematic investigation of the potential moderating roles of sex and immigration background.
Limitations
The first limitation is that none of the two studies relied on a longitudinal design. As such, the hypothetical developmental trends identified in this study could also be the result of a cohort effect (Little 2013) and that reciprocal effects of engagement on behavior problems might also be at play. Moreover, engagement tends to remain relatively stable over time both in childhood (Archambault and Dupéré 2016) and adolescence (Wang and Eccles 2012). Yet, it was not possible to control for preexisting levels of engagement, which would have strengthened the results.
Second, because this investigation combines two independent samples of participants initially collected for different purposes, the study had to rely on measures that were not exactly equivalent in the two samples despite being validated for each developmental period. As typically done in studies assessing externalizing behaviors and student engagement (see Table 1), teachers’ ratings of students’ externalizing behaviors were obtained in the elementary school sample, and students’ self-reports of these same behaviors were used in the secondary school sample. This decision is consistent with the greater ease with which elementary teachers can report on the externalizing behaviors of children with whom they interact all day, five days a week (Smith 2007). In contrast, secondary school teachers only have limited contact with their students, typically limited to a few hours of in-class teaching, making it virtually impossible for them to report reliably on the whole range of externalizing behaviors exhibited by their students (Randazzo et al. 2003). Conversely, the measure of internalizing behaviors was self-reported in both samples. Given the naturally internal and hard-to-observed nature of these behaviors, youth are considered more reliable informants (Smith 2007). Also, the measure was longer in the secondary school sample than in the elementary school sample. Respectively, each measure was selected to reflect best students’ developmental stage, as well as the normative attention span of youth from this age group (which is typically longer in adolescents than in children; Fuchs 2005). Studies have shown that different questionnaires measuring internalizing behaviors can function equally well even if one is more extensive than the other (Dart et al. 2020).
The student engagement measure was also not equivalent across samples, as the two measures used have only been validated in their respective age group (elementary or secondary school samples). The elementary school measure focused on language class, whereas the secondary school measure encompassed students’ more general school engagement. Although the emotional dimension covered relatively similar emotions but did so more extensively in the secondary school sample, the behavioral dimension had a stronger focus on compliance among elementary students, and disruptiveness among secondary students. Conversely, the cognitive dimension was more extensively covered in secondary students, whose self-regulation abilities tend to be more developed than those of elementary students. These two dimensions are thus consistent with known developmental differences in relation to student engagement between younger (Archambault and Dupéré 2016) and older students (Wang and Eccles 2012). As for the internalizing behavior measure, the longer length of the secondary school measure of engagement aimed to match adolescents’ greater attention span. However, even if all measures were selected to best represent the developmental stage of the students, it remains that the reliance on distinct measures may complicate the comparison of results across studies, especially with the specific factors. Thus, this limitation calls for replication with measures that are perhaps more comparable across developmental periods. Still, these differences in measures also reinforce the robustness of the results that are common to both studies.
Third, considering the number of parameters included in the models, the complexity of bifactor models, and the sample size, it was impossible to rely on a fully latent approach. To overcome this limitation, the study relied on factor scores, which still control for a part of measurement error (DiStefano et al. 2009). Finally, neither this investigation nor the two studies that provided the data used in the present investigation aimed to assess ethnic or immigration status differences. For this reason, one sample was relatively homogenous (elementary school), whereas the other only provided limited information about the students’ immigration background (secondary school). This information allowed conducting tests of measurement and predictive invariance as a function of immigration status among secondary school students and controlling for this variable as part of the main analyses of this study. These verifications revealed very few differences, consistent with the robustness of the findings across distinct immigration groups. Still, due to the very limited amount of information available in this regard, this study cannot be considered to represent a systematic assessment of immigration differences, and additional studies relying on more diverse samples of children and adolescents will be required to understand better the role played by culture, ethnicity, and nationality in these relations.
Conclusion
This study formally assessed the joint risk posed by global and specific levels of externalizing and internalizing behavior problems for student engagement. The negative role of externalizing behavior on students’ global levels of engagement and their specific levels of behavioral engagement remained salient in elementary and secondary school students. In contrast, age differences were found pertaining to internalizing behaviors. Among elementary school students, specific levels of anxiety, and global levels of internalizing behaviors were important risk factors for reduced engagement (global, emotional, and cognitive). Among secondary school students, specific levels of depressive symptoms became more prominent risk factors for global and emotional engagement. Despite these possible developmental effects, the results highlight that externalizing and internalizing behaviors represent risk factors for school disengagement equally among boys and girls, as well as immigrants and non-immigrants. Moreover, the results do not support that students who accumulate behavior problems are more at risk of disengagement than the risk already posed by displaying any one type of problem. These results generally suggest that prevention strategies targeting externalizing behaviors should be favored before and after the transition to secondary school. In contrast, prevention strategies designed to minimize the impact of internalizing behaviors could focus on anxiety and global levels of internalizing behaviors among elementary students and on depressive symptoms following the transition to secondary school.
Notes
The literature review was conducted using the PsycInfo and Web of Science search engines as well (keywords details: (anxi* or depress* or internali* or hyperac* or inatten* or aggress* or defian* or opposit* or disrupt* or externali*) and (engag* or motiva*) and (school or student or classroom or acad*) which were restricted to the title and peer reviewed articles). The search yield a total of 234 article. Of those, 63 were retained based on the title, and 20 matched the inclusion criterial, namely to assess at least one behavior problem in association student engagement or a closely related measure (motivation achievement goals (n = 4), pre-school learning skills (n = 3), and school connectedness (n = 1) are excluded from Table 1, but discussed in the text when appropriate). Studies that assessed a subject-specific adjustment problem (e.g., mathematic anxiety) or focusing solely on delinquency involving illegal actions (e.g., bringing drugs to school) or outside of school delinquency (e.g., gang participation) were also excluded from Table 1. Screening the reference list of the retained article (i.e., forward snowballing) yield an additional 3 studies also reported in Table 1.
See Appendix 2 for a detailed description of the measurement models.
The original scale included a fifth item (“This student steals from home, school, or elsewhere.”). However, because of the rarity of this behavior in elementary school children and the difficulty for teachers to observe “hidden” behaviors, this item was not included in this study.
See Appendix 2 for a detailed description of the measurement models.
References
Achenbach, T. M., & Edelbrock, C. S. (1978). The classification of child psychopathology: a review and analysis of empirical efforts. Psychological Bulletin, 85(6), 1275–1301.
Al-Hendawi, M. (2012). Academic engagement of students with emotional and behavioral disorders: existing research, issues, and future directions. Emotional and Behavioural Difficulties, 17(2), 125–141.
Archambault, I., & Dupéré, V. (2016). Joint trajectories of behavioral, affective, and cognitive engagement in elementary school. The Journal of Educational Research, 110(2), 188–198.
Archambault, I., Janosz, M., Fallu, J. S., & Pagani, L. S. (2009). Student engagement and its relationship with early high school dropout. Journal of Adolescence, 32(3), 651–670.
Archambault, I., Janosz, M., Morizot, J., & Pagani, L. S. (2009). Adolescent behavioral, affective, and cognitive engagement in school: relationship to dropout. Journal of School Health, 79(9), 408–415.
Archambault, I., & Vandenbossche-Makombo, J. (2014). Validation de l’échelle des dimensions de l’engagement scolaire (ÉDES) chez les élèves du primaire. Canadian Journal of Behavioural Science/Revue canadienne des sciences du comportement, 46(2), 275–288.
Archambault, I., Vandenbossche-Makombo, J., & Fraser, S. L. (2017). Students’ oppositional behaviors and engagement in school: the differential role of the student-teacher relationship. Journal of Child and Family Studies, 26(6), 1702–1712.
Baker, J. A., Grant, S., & Morlock, L. (2008). The teacher-student relationship as a developmental context for children with internalizing or externalizing behaviors. School Psychology Quarterly, 23(1), 3–15.
Bierman, K. L., & Sasser, T. R. (2014). Conduct disorder. In M. Lewis & K. D. Rudolph (Eds.), Handbook of developmental psychopathology, 3rd Ed. (pp. 467–487). New York, NY: Springer.
Birmaher, B., Brent, D. A., Chiappetta, L., Bridge, J., Monga, S., & Baugher, M. (1999). Psychometric properties of the Screen fo Child Anxiety Related Emotional Disorders (SCARED): a replication study. Journal of the American Academic Child and Adolescent Psychiatry, 38(10), 1230–1236.
Blais, M., Hess, U., & Riddle, A. S. (2002). A multi-group investigation of the CES-D’s measurement structure across adolescents, young adults and middle-aged adults. CIRANO Working Papers, 2002s-36.
Caci, H., Morin, A. J., & Tran, A. (2015). Investigation of a bifactor model of the Strengths and Difficulties Questionnaire. European Child and Adolescent Psychiatry, 24(10), 1291–1301.
Campbell, S. B., Halperin, J. M., & Sonuga-Barke, E. J. S. (2014). A developmental perspective on attention-deficit/hyperactivity disorder (ADHD). In M. Lewis & K. D. Rudolph (Eds.), Handbook of developmental psychopathology, 3rd Ed. (pp. 427–448). New York, NY: Springer.
Capron, C., Thérond, C., & Duyme, M. (2007). Psychometric properties of the French version of the self-report and teacher Strengths and Difficulties Questionnaire (SDQ). European Journal of Psychological Assessment, 23(2), 79–88.
Caspi, A., Houts, R. M., Belsky, D. W., Goldman-Mellor, S. J., Harrington, H., Israel, S., Meirer, M. H., Ramrakha, S., Shalev, I., Poulton, R., & Moffitt, T. E. (2014). The p factor: one general psychopathology factor in the structure of psychiatric disorders? Clinical Psychological Science, 2(2), 119–137.
Chen, F. F. (2007). Sensitivity of goodness of fit indexes to lack of measurement invariance. Structural Equation Modeling: A Multidisciplinary Journal, 14(3), 464–504.
Curhan, A. L., Rabinowitz, J. A., Pas, E. T., & Bradshaw, C. P. (2020). Informant discrepancies in internalizing and externalizing symptoms in an at-risk sample: the role of parenting and school engagement. Journal of Youth and Adolescence, 49(1), 311–322.
Dart, E. H., Arora, P., Collins, T. A., Stark, K., Cook, C. R., Duong, M. T., McCarthy, C. A., & Doll, B. (2020). Developing a brief behavior rating scale for progress monitoring of depression in school settings. Assessment for Effective Intervention, 45(3), 163–172.
Demaray, M. K., & Jenkins, L. N. (2011). Relations among academic enablers and academic achievement in children with and without high levels of parent-rated symptoms of inattention, impulsivity, and hyperactivity. Psychology in the Schools, 48(6), 573–586.
Derdikman-Eiron, R., Indredavik, M. S., Bratberg, G. H., Taraldsen, G., Bakken, I. J., & Colton, M. (2011). Gender differences in subjective well-being, self-esteem and psychosocial functioning in adolescents with symptoms of anxiety and depression: findings from the Nord-Trondelag health study. Scandinavian Journal of Psychology, 52(3), 261–267.
DiStefano, C., Zhu, M., & Mindrila, D. (2009). Understanding and using factor scores: considerations for the applied researcher. Practical Assessment, Research & Evaluation, 14(20), 1–11.
Dorio, N. B., Secord Fredrick, S., & Demaray, M. K. (2019). School engagement and the role of peer victimization, depressive symptoms, and rumination. Journal of Early Adolescence, 39(7), 962–992.
Duncan, G. J., Dowsett, C. J., Claessens, A., Magnuson, K., Huston, A. C., Klebanov, P., Pagani, L. S., Feinstein, L., Engel, M., Brooks-Gunn, J., Sexton, H., Duckworth, K., & Japel, C. (2007). School readiness and achievement. Developmental Psychology, 43(6), 1428–1446.
Eccles, J. S. & Roeser, R. W. (2009). Schools, academic motivation, and stage-environment fit. In R. M. Lerner & L. Steinberg (Eds.). Handbook of adolescent psychology (pp. 404–434). Hoboken, NJ: Wiley.
Eisenberg, N., Valiente, C., Spinrad, T. L., Liew, J., Zhou, Q., Losoya, S. H., Reiser, M., & Cumberland, A. (2009). Longitudinal relations of children’s effortful control, impulsivity, and negative emotionality to their externalizing, internalizing, and co-occurring behavior problems. Developmental Psychology, 45(4), 988–1008.
Enders, C. K. (2010). Applied missing data analysis. New York, NY: Guildford.
Evans, G. W., Li, D., & Whipple, S. S. (2013). Cumulative risk and child development. Psychological Bulletin, 139(6), 1342–1396.
Fiorilli, C., De Stasio, S., Di Chiacchio, C., Pepe, A., & Salmela-Aro, K. (2017). School burnout, depressive symptoms and engagement: their combined effect on student achievement. International Journal of Educational Research, 84, 1–12.
Flake, J. K., & Petway, K. R. (2019). Methodologies for investigating and interpreting student-teacher rating incongruence in noncognitive assessment. Educational Measurement: Issues and Practice, 38(1), 63–77.
Fredricks, J. A., Blumenfeld, P. C., & Paris, A. H. (2004). School engagement: potential of the concept, state of the evidence. Review of Educational Research, 74(1), 59–109.
Fredricks, J. A., Filsecker, M., & Lawson, M. (2016). Student engagement, context, and adjustment: addressing definitional, measurement, and method issues. Learning and Instruction, 43, 1–4.
Fuchs, M. (2005). Children and adolescents as respondents. Journal of Official Statistics, 21, 701–725.
Garber, J., & Rao, U. (2014). Depression in children and adolescents. In M. Lewis & K. D. Rudolph (Eds.), Handbook of developmental psychopathology, 3rd Ed. (pp. 489–520). New York, NY: Springer.
Garvik, M., Idsoe, T., & Bru, E. (2014). Depression and school engagement among Norwegian secondary vocational school students. Scandinavian Journal of Educational Research, 58(5), 592–608.
Gillen-O’Neel, C. G., Ruble, D. N., & Fuligni, A. J. (2011). Ethnic stigma, academic anxiety, and intrinsic motivation in middle childhood. Child Developement, 82(5), 1470–1485.
Gogol, K., Brunner, M., Martin, R., Preckel, F., & Goetz, T. (2017). Affect and motivation within and between school subjects: development and validation of an integrative structural model of academic self-concept, interest, and anxiety. Contemporary Educational Psychology, 49, 46–65.
Goodman, R. (2001). Psychometric properties of the Strengths and Difficulties Questionnaire. Journal of the American Academy of Child and Adolescent Psychiatry, 40(11), 1337–1345.
Gut, J., Heckmann, C., Meyer, C. S., Schmid, M., & Grob, A. (2012). Language skills, mathematical thinking, and achievement motivation in children with ADHD, disruptive disorders, and normal controls. Learning and Individual Differences, 22(3), 375–379.
Hoge, R. D., Meginbir, L., Khan, Y., & Weatherall, D. (1985). A multitrait-multimethod analysis of the Preschool Behavior Questionnaire. Journal of Abnormal Child Psychology, 13, 119–127.
Jarrett, M., Siddiqui, S., Lochman, J., & Qu, L. (2014). Internalizing problems as a predictor of change in externalizing problems in at-risk youth. Journal of Clinical Child & Adolescent Psychology, 43(1), 27–35.
Junod, R. E. V., DuPaul, G. J., Jitendra, A. K., Volpe, R. J., & Cleary, K. S. (2006). Classroom observations of students with and without ADHD. Journal of School Psychology, 44(2), 87–104.
Konold, T., Cornell, D., Shukla, K., & Huang, F. (2017). Racial/ethnic differences in perceptions of school climate and its association with student engagement and peer aggression. Journal of Youth and Adolescence, 46(6), 1289–1303.
Kurdi, V., & Archambault, I. (2020). Self-perceptions and engagement in low socio-economic elementary school students: the moderating effects of immigration status and anxiety. School Mental Health, 12, 400–416.
Li, Y., Bebiroglu, N., Phelps, E., Lerner, J. V., & Lerner, R. M. (2008). Out-of-school time activity participation, school engagement and positive youth development: findings from the 4-H Study of Positive Youth Development. Journal of Youth Development, 3(3), 1–16.
Li, Y., & Lerner, R. M. (2011). Trajectories of school engagement during adolescence: implications for grades, depression, delinquency, and substance use. Developmental Psychology, 47(1), 233–247.
Little, T. D. (2013). Longitudinal structural equation modeling. New York, NY: The Guildford Press.
Marsh, H. W., Hau, K. T., & Grayson, D. (2005). Goodness of fit in structural equation models. In A. Maydeu-Olivares & J. J. McArdle (Eds.), Contemporary psychometrics (pp. 275–340). Mahwah: NJ: Erlbaum.
Marsh, H. W., Hau, K. T., Wen, Z., Nagengast, B., & Morin, A. J. S. (2013). Moderation. In T. D. Little (Ed.), Oxford handbook of quantitative methods. New York, NY: Oxford University Press.
MEES. (2019). Diplomation et qualification par commission scolaire au secondaire. Québec, Qc: Ministère de l’Éducation et de l’Enseignement supérieur du Québec.
Moilanen, K. L., Shaw, D. S., & Maxwell, K. L. (2010). Developmental cascades: externalizing, internalizing, and academic competence from middle childhood to early adolescence. Development and Psychopathology, 22(3), 635–653.
Morin, A. J. S., Arens, A. K., Maiano, C., Ciarrochi, J., Tracey, D., Parker, P. D., & Craven, R. G. (2017). Reciprocal relationships between teacher ratings of internalizing and externalizing behaviors in adolescents with different levels of cognitive abilities. Journal of Youth & Adolescence, 46(4), 801–825.
Morin, A. J. S., Boudrias, J. S., Marsh, H. W., Madore, I., & Desrumaux, P. (2016). Further reflections on disentangling shape and level effects in person-centered analyses: an illustration exploring the dimensionality of psychological health. Structural Equation Modeling, 23(3), 438–454.
Muthén, L. K., & Muthén, B. O. (2019). Mplus user’s guide. Los Angeles, CA: Muthén & Muthén.
Nguyen, M. N., Watanabe-Galloway, S., Hill, J. L., Siahpush, M., Tibbits, M. K., & Wichman, C. (2019). Ecological model of school engagement and attention-deficit/hyperactivity disorder in school-aged children. European Child and Adolescent Psychiatry, 28(6), 795–805.
Nivard, M. G., Lubke, G. H., Dolan, C. V., Evans, D. M., St Pourcain, B., Munafo, M., & Middeldorp, C. M. (2017). Joint developmental trajectories of internalizing and externalizing disorders between childhood and adolescence. Development and Psychopathology, 29(3), 919–928.
OECD. (2018). Equity in education: breaking down barriers to social mobility. Paris: PISA, OECD.
Olivier, E., & Archambault, I. (2017). Hyperactivity, inattention, and student engagement: the protective role of relationships with teachers and peers. Learning and Individual Differences, 59, 86–95.
Olivier, E., Archambault, I., & Dupéré, V. (2018). Boys’ and girls’ latent profiles of behavior and social adjustment in school: longitudinal links with later student behavioral engagement and academic achievement? Journal of School Psychology, 69, 28–44.
Olivier, E., Archambault, I., & Dupéré, V. (2020). Do needs for competence and relatedness mediate the risk of low engagement of students with behavior and social problem profiles? Learning & Individual Differences, 78, No pagination specified.
Papachristou, E., & Flouri, E. (2019). The codevelopment of internalizing symptoms, externalizing symptoms, and cognitive ability across childhood and adolescence. Development and Psychopathology, 7, 1–15.
Plamondon, A., & Martinussen, R. (2019). Inattention symptoms are associated with academic achievement mostly through variance shared with intrinsic motivation and behavioral engagement. Journal of Attention Disorders, 23(14), 1816–1828.
Radloff, L. S. (1977). The CES-D scale: a self-report depression scale for research in the general population. Applied Psychological Measurement, 1, 385–401.
Randazzo, K. V., Landsverk, J., & Ganger, W. (2003). Three informants’ reports of child behavior: parents, teachers, and foster parents. Journal of the American Academy of Child and Adolescent Psychiatry, 42(11), 1343–1350.
Reitz, E., Deković, M., & Meijer, A. M. (2005). The structure and stability of externalizing and internalizing problem behavior during early adolescence. Journal of Youth and Adolescence, 34(6), 577–588.
Searle, A. K., Miller-Lewis, L. R., Sawyer, M. G., & Baghurst, P. A. (2013). Predictors of children’s kindergarten classroom engagement: preschool adult–child relationships, self-concept, and hyperactivity/inattention. Early Education & Development, 24(8), 1112–1136.
Skinner, E. A., Kindermann, T. A., Connell, J. P., & Wellborn, J. G. (2009). Engagement and disaffection as organizational construct in the dynamics of motivational development. In K. R. Wentzel & A. Wigfield (Eds.), Handbook of motivation at school (pp. 223–246). New York, NY: Routledge.
Smith, S. R. (2007). Making sense of multiple informants in child and adolescent psychopathology: a guide for clinicians. Journal of Psychoeducational Assessment, 25(2), 139–149.
Tardif-Grenier, K., Olivier, É., & Archambaul, I. (2019). Is there an immigrant paradox in Canadian elementary students’ behavioral and social adjustment? International Journal of School & Educational Psychology. https://doi.org/10.1080/21683603.2019.1631237.
Tian, Y., Bian, Y., Han, P., Gao, F., & Wang, P. (2017). Class collective efficacy and class size as moderators of the relationship between junior middle school students’ externalizing behavior and academic engagement: a multilevel study. Frontiers in Psychology, 8, 1–9.
Trembaly, R. E., Desmarais-Gervais, L., Gagnon, C., & Charlebois, P. (1987). The Preschool Behavior Questionnaire: stability of its factor structure between cultures, sexes, ages and socioecconomic classes. Internationnal Journal of Behavioral Development, 10, 467–484.
UNESCO. (2020). World inequality database on education. Montréal, QC: UNESCO Institute for Statistics.
Vasey, M., Bosmans, G., & Ollendick, T. (2014). The developmental psychopathology of anxiety. In M. Lewis & K. Rudolph (Eds.), Handbook of developmental psychopathology. 3rd Ed. (pp. 543–560). New York, NY: Springer.
Volpe, R. J., DuPaul, G. J., DiPerna, J. C., Jitendra, A. K., Lutz, J. G., Tresco, K., & Junod, R. E. V. (2006). Attention deficit hyperactivity disorder and scholastic achievement: A model of mediation via academic enablers. School Psychology Review, 35(1), 47–61.
Wang, F. L., Eisenberg, N., Valiente, C., & Spinrad, T. L. (2016a). Role of temperament in early adolescent pure and co-occurring internalizing and externalizing problems using a bifactor model. Development and Psychopathology, 28(4pt2), 1487–1504.
Wang, M.-T., Chow, A., Hofkens, T., & Salmela-Aro, K. (2015). The trajectories of student emotional engagement and school burnout with academic and psychological development: findings from Finnish adolescents. Learning and Instruction, 36, 57–65.
Wang, M. T., & Eccles, J. S. (2012). Adolescent behavioral, emotional, and cognitive engagement trajectories in school and their differential relations to educational success. Journal of Research on Adolescence, 22(1), 31–39.
Wang, M. T., Fredricks, J., Ye, F., Hofkens, T., & Linn, J. S. (2019). Conceptualization and assessment of adolescents’ engagement and disengagement in school. European Journal of Psychological Assessment, 35(4), 592–606.
Wang, M. T., & Fredricks, J. (2014). The reciprocal links between school engagement, youth problem behaviors, and school dropout during adolescence. Child Developement, 85(2), 722–737.
Wang, M. T., Fredricks, J. A., Ye, F., Hofkens, T. L., & Linn, J. S. (2016b). The Math and Science Engagement Scales: scale development, validation, and psychometric properties. Learning and Instruction, 43, 16–26.
Wang, M. T., & Peck, S. C. (2013). Adolescent educational success and mental health vary across school engagement profiles. Developmental Psychology, 49(7), 1266–1276.
Wigfield, A., & Eccles, J. S. (2000). Expectancy-value theory of achievement motivation. Contemporary Educational Psychology, 25(1), 68–81.
Willner, C. J., Gatzke-Kopp, L. M., & Bray, B. C. (2016). The dynamics of internalizing and externalizing comorbidity across the early school years. Development and Psychopathology, 28(4pt1), 1033–1052.
Acknowledgements
The first author was supported by funding from the Canadian Institute for Military and Veteran Health Research (CIMVHR). The second author was supported by a grant from the Social Sciences and Humanities Research Council of Canada (435-2018-0368). Data collection was supported by grants from the Fonds de recherche Société et culture (elementary school project: #131430; secondary school project: 2018-NP-205005) and from the Social Sciences and Humanities Research Council of Canada (secondary school project; 430-2018-00456). The fifth author was also supported by funding from the Canada Research Chairs program.
Author information
Authors and Affiliations
Contributions
E.O. conceived the study, analyzed and interpreted the data, and wrote the original and revised manuscript; A.J.S.M. conceived the study, analyzed and interpreted the data, and wrote the original and revised manuscript; J.L. participated to the literature review and to writing the original and revised manuscript; K.T.G. provided the data and participated to writing the original and revised manuscript; I.A. provided the data and participated to writing the original and revised manuscript. All authors read and approved the final manuscript.
Data Sharing and Declaration
The datasets analyzed during the current study are not publicly available but are available from the corresponding author on reasonable request and with permission of I.A. and K.T.-G.
Corresponding author
Ethics declarations
Conflict of Interest
The authors declare that they have no conflict of interest.
Ethical Approval
All procedures performed in this study were approved by the ethics committee of the University of Montreal. This research complies with APA’s ethical standards in the treatment of human samples and with the highest ethical standards.
Informed Consent
Active informed consent was obtained from all individual participants included in the study, both teachers and students. Parents also gave their active consent for their child to participate.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary information
Rights and permissions
About this article
Cite this article
Olivier, E., Morin, A.J.S., Langlois, J. et al. Internalizing and Externalizing Behavior Problems and Student Engagement in Elementary and Secondary School Students. J Youth Adolescence 49, 2327–2346 (2020). https://doi.org/10.1007/s10964-020-01295-x
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10964-020-01295-x