Search

Support from parents and school staff are important for adolescent well-being. However, few studies have examined the role of parental involvement and adult social support at school on high school students' mental health, specifically, suicidal thoughts and behaviors (STBs). Data were collected from 362,980 high school students (51.85% female) from 412 schools in Georgia. Students reported perceptions of parental involvement, adult social support at school, experiences of face-to-face victimization, cyber victimization, and STBs. Hierarchical linear modeling (HLM) results showed that perceived parental involvement and adult social support at the student and school levels were negatively related to adolescent STBs. Face-to-face victimization and cyber victimization at the student and school levels were significant predictors for STBs. Student-level perceived parental involvement and adult social support mitigated the relationship between peer victimization and STBs. However, school-level perceived adult social support exacerbated the relationship between face-to-face and cyber victimization and STBs. It is essential to attend to the differential effect of student-level and school-level adult social support on youth mental health. It is also important to promote parental involvement and adult social support to promote mental health among high school students.

In this study, we investigate how the level of work control predictability affects employee creativity. Specifically, we examine whether supervisor and coworker support moderate the predictability-creativity relationship. We use survey data from 128 employee--supervisor dyads from a governmental organization in Belgium. Multilevel analyses demonstrated a significant inverted U-curve relationship between predictability and leader-rated creativity, indicating that too little or too much work control predictability contributes less to employee creativity compared to moderate levels of work control predictability. Moreover, supervisor and coworker support moderate this relationship in such a way that for employees who receive low support the inverted U-curve relationship is significantly stronger compared to employees that receive high support. This is the first study that focuses on work control predictability in relation to creative performance.

This study analyzed the latest four PISA surveys, 2009, 2012, 2015, and 2018, to explore the association between students' ICT-related use and math and science performance. Using ICT Engagement Theory as a theoretical framework and a three-level hierarchical linear modeling approach, while controlling for confounding effects, ICT-related independent variables of interest were added to the models at the student, school, and country levels. The series of models revealed that, in general, an increase in ICT availability and ICT use both inside and outside school had a negative association with learning outcomes, while students' positive attitude toward ICT demonstrated a strong positive relationship. However, students' perceived autonomy related to ICT use had the strongest association with academic performance, which is consistent with the changing nature of the modern learning environments. Findings revealed that virtually all forms of student ICT use, both inside and outside of school and whether subject related or not, had no substantive positive relationship with student performance in math or science. Conversely, higher student attitude toward, confidence in, belief in utility of, and autonomous use of ICT was associated with higher math and science performance for each of the four years of the study. Incidentally, we also found that while country GDP per capita had no consistent association with student performance, a school's provision of extra-curricula activities did. Recommendations for educational leaders, teachers, and parents are offered.

Learning processes often occur at a situational level. Changes in learning context have implications on how students are motivated or are able to cognitively process information. To study such situational phenomena, Experience Sampling Method (ESM) can help assess psychological variables in the moment and in context. However, data collected via ESM is voluminous and imbalanced. Special types of statistical modeling are needed to handle this unique data structure in order to maximize its potential for scientific discovery. The purpose of this paper is to illustrate how Latent State-Trait modeling used within a multilevel framework can help model complex data as derived by ESM. A study of situational engagement is presented as an illustrative case. We describe methodological considerations which facilitated the following analyses: (1) "Decomposition of trait-level and state-level engagement"; (2) "Group differences in variance decomposition"; and (3) "Predicting state component of engagement." Discussions include the relative advantages and disadvantages of ESM and multilevel Latent State-Trait modeling in facilitating situational psychological research.

Intensive longitudinal data is increasingly used to study state-like processes such as changes in daily stress. Measures aimed at collecting such data require the same level of scrutiny regarding scale reliability as traditional questionnaires. The most prevalent methods used to assess reliability of intensive longitudinal measures are based on the generalizability theory or a multilevel factor analytic approach. However, the application of recent improvements made for the factor analytic approach may not be readily applicable for all researchers. Therefore, this article illustrates a five-step approach for determining reliability of daily data, which is one type of intensive longitudinal data. First, we show how the proposed reliability equations are applied. Next, we illustrate how these equations are used as part of our five-step approach with empirical data, originating from a study investigating changes in daily stress of secondary school teachers. The results are a within-level ([omega] superscript w), between-level ([omega] superscript b) reliability score. Mplus syntax for these examples is included and discussed. As such, this paper anticipates on the need for comprehensive guides for the analysis of daily data.

Multilevel modelling (MLM) is the most frequently used approach for evaluating interventions with clustered data. MLM, however, has some limitations that are associated with numerous obstacles to model estimation and valid inferences. Longitudinal multiple-group (LMG) modelling is a longstanding approach for testing intervention effects using cluster-sampled data that has been superseded by the rise of MLM approaches, but the LMG approach can have advantages when research questions do not pertain to predicting variability at the higher levels. In this paper, we first review the advantages and limitations of MLM and LMG approaches. Second, steps in the estimation of an LMG model are presented, with some recent upgrades and changes in the modelling strategy that have particular utility for evaluating interventions. We discuss the advantages of the LMG approach as a guided confirmatory model-testing framework and how the approach places a premium on avoiding Type II errors, particularly when complex interactions are potentially at play.

Educational disadvantage is a multilevel and multifactorial phenomenon, with a myriad of contributing factors having been identified within the social contexts inhabited by the student from a lower socioeconomic status (SES) background. Irish research has accumulated a comprehensive body of empirical evidence on the factors operating within family, school and societal contexts but less attention has been paid to the role that the neighbourhood might have in contributing to the under-achievement of students from lower SES backgrounds. This study employed data from the 9 and 17/18-year-old cohorts of the Growing Up in Ireland survey to explore whether the SES of the neighbourhood within which a student lives makes a significant contribution to the prediction of academic achievement over and above the SES of the family and school. A series of multilevel models demonstrated the additive effects of social contexts on academic achievement in both samples. Each social context provided a unique, statistically significant contribution to the prediction of academic achievement, with smaller contributions made by more distal contexts. These findings support an understanding of educational disadvantage as a phenomenon that is perpetuated within multiple social contexts and suggest that strategies to tackle educational disadvantage should take a pan-contextual perspective.

In education, multisite trials involve randomization of pupils into intervention and comparison groups within schools. Most analytical models in multisite educational trials ignore that the impact of an intervention may be school dependent. This study investigates the impact of statistical models on the uncertainty associated with an effect size using comparable outcomes and covariates from ten multisite educational trials funded by the UK's Education Endowment Foundation. Ordinary least squares (OLS) models often assume that the pupil's outcomes within schools are independent, which is not always true. Multilevel models address this limitation by incorporating heterogeneity between schools to account for intra-school dependency. This inflates the confidence interval of an effect size obtained from the multilevel models than from an OLS model. For a multisite trial, the heterogeneity between schools also includes the differences in the expected impact of intervention between schools. Ignoring this additional school-by-intervention variation in a multisite trial could affect both its interpretation and conclusions. A robust approach to estimate the confidence intervals for effect size from multisite trials is by treating effect size as a parameter with its distribution. This paper is important for evaluating evidence from multisite trials by accounting for all sources of variability.

Background: Given the relatively low graduation and retention rate in undergraduate engineering programs in the United States, the factors that influence student success outcomes need to be examined. However, limited research systematically studied both student- and school-level factors and how they influenced undergraduate engineering student success outcomes. We gathered responses from 458 engineering undergraduate students in a cross-sectional multilevel multi-school (14 schools) survey. These 14 schools included both large state universities and liberal arts colleges. The survey measured various student-level factors, including demographic, skills, and personality variables, along with seven school-level factors, such as student-faculty ratio and school type (i.e., public versus private). The data were analyzed using the hierarchical multilevel modeling approach. Results: The results showed that female students reported better outcomes than male students, racial minority students reported better outcomes than White students, but first-generation students reported poorer outcomes. Communication competency was associated with student learning outcomes, GPA, and program satisfaction, whereas conflict management preferences were not significantly correlated with any student success outcomes. The results of the school-level factors' influences on student success outcomes were not consistent, but some factors, such as student-faculty ratios and diversity rate, were significantly related to some student outcomes. Conclusion: Engineering education is a complex, multi-faceted issue that requires more collaborative and systematic research. We hope our findings help educators understand the different factors that could potentially influence engineering students and inform better program design and policymaking.

Multilevel latent class analysis (MLCA) has been increasingly used to investigate unobserved population heterogeneity while taking into account data dependency. Nonparametric MLCA has gained much popularity due to the advantage of classifying both individuals and clusters into latent classes. This study demonstrated the need to relax the assumption in specifying the nonparametric MLCA: item response probabilities varied only across level-1 latent classes, but not level-2 latent classes. An empirical demonstration with data from the Trends in International Mathematics and Science Study (TIMSS) 2011 showed that item response probabilities could vary across both level-1 and level-2 latent classes. This relaxed MLCA yielded better model fit and provided more nuanced understanding of the heterogeneous response patterns. Monte Carlo simulation was conducted to evaluate class enumeration and assignment accuracy of the relaxed MLCA. Based on the simulation results, we recommended the use of AIC in class enumeration and highlighted the benefits of having larger cluster size.

Decomposing variables into between and within components are often required in multilevel analysis. This method of decomposition should not ignore possible unreliability of an observed group mean (i.e., arithmetic mean) that is due to small cluster sizes and can lead to substantially biased estimates. Adjustment procedures that allow unbiased estimation have been defined and implemented in software for a two-level model. This study shows how to implement a two-stage adjustment procedure in a three-level design. A simulation study showed that the adjustment procedure provides unbiased estimates. To demonstrate how the adjustment procedure can change results in a real data context, an illustration is provided using a set up in which 355 Level-1 units are nested in 93 Level-2 and 19 Level-3 units.

Large-scale education data are collected via complex sampling designs that incorporate clustering and unequal probability of selection. Multilevel models are often utilized to account for clustering effects. The probability weighted approach (PWA) has been frequently used to deal with the unequal probability of selection. In this study, we examine the performance of an intuitive, easy to implement approach named the sample distribution approach (SDA) that utilizes Markov Chain Monte Carlo (MCMC) methods and Bayesian inference. Our simulation design focused on clustering effects, represented by the Intraclass Correlation (ICC) and on the sample size of the cluster. We analyzed a large-scale educational assessment dataset (Early Childhood Longitudinal Study -- Kindergarten 2011) to compute estimates for the simulation. Findings reveal that the SDA overall generated reliable posterior distributions of parameters and had small error variances. In addition, although design informativeness is important, the ICC and cluster sample size factors had little impact on the performance of this model-based approach.

In line with assumptions made by the control-value theory of academic emotions, it was hypothesized that the association between the classroom social environment, in terms of students' perceptions of their teachers' interpersonal behaviour, and students' academic emotions was partially mediated by students' achievement goals. The present study applied multilevel structural equation modelling on questionnaire data from a sample of 2000 Chinese secondary school students. As assumed, achievement goals partially mediated the association between the classroom social environment and student emotions. Further, results indicated that the classroom social environment was more closely associated with student emotions than student goals were. The findings of this study add to the understanding of the antecedents of students' emotions in class.

Multivariate analysis of variance (MANOVA) is widely used to test the null hypothesis of equal multivariate means across 2 or more groups. MANOVA rests upon an assumption that error terms are independent of one another, which can be violated if individuals are clustered or nested within groups, such as schools. Ignoring such nesting can result in Type I error inflation, biased parameter estimates, and underestimated standard errors. This simulation study compared two approaches for testing the MANOVA null hypothesis of no group mean differences, when data come from a multilevel structure, under a variety of conditions. Results indicated that the multilevel MANOVA method of Snijders and Bosker, as well as an approach based on multilevel structural equation modeling (SEM) controlled Type I error under most conditions. In addition, the SEM based method yielded slightly higher power than did the multilevel MANOVA, under some conditions. Implications for practice are discussed.

Longitudinal data structures are frequently encountered in a variety of disciplines in the social and behavioral sciences. Growth curve modeling offers a highly extensible framework that allows for the exploration of rich hypotheses. However, owing to the presence of interrelated sources of potential data-model misfit at multiple levels, the matter of model criticism remains challenging for even foundational growth curve models. Through a simulation study and an applied example, the performance of six discrepancy measures was investigated using posterior predictive model checking as the framework for model criticism. The likelihood ratio and the standardized generalized dimensionality discrepancy measure outperformed the other discrepancy measures under consideration and show promise for future study and use.

Optimal sampling frameworks attempt to identify the most efficient sampling plans to achieve an adequate statistical power. Although such calculations are theoretical in nature, they are critical to the judicious and wise use of funding because they serve as important starting points that guide practical discussions around sampling tradeoffs and requirements. Conventional optimal sampling frameworks, however, often identify sub-optimal designs because they typically presume the costs of sampling units are equal across treatment conditions. In this study, we develop a more flexible framework that allows costs to differ by treatment conditions and derive the optimal sample size formulas for three-level multisite cluster-randomized trials. We find that the proposed optimal sampling schemes are driven by the differences in costs between treatment conditions, cross-level sampling cost ratios and cross-level variance decomposition ratios. We illustrate the utility of the proposed framework by comparing it to a conventional framework and find that the proposed framework frequently identifies more efficient designs. The proposed optimal sampling framework has been implemented in the "R" package "odr."

We examined the evolution of social pretend play in toddlers and tested whether children's age, gender, and language abilities were associated with changes in social pretend play over time. We employed a unique observational data set that followed 28 children (M[subscript age] = 27 months old; 57% boys) with each of two unfamiliar peers over 18 playdates per dyad. Cross-classified multilevel models revealed a quadratic change in the number of successful initiations in toddlers' social pretend play over time with a more pronounced curve for older toddlers. Boys' failed initiations increased linearly over time, whereas girls' failed initiations remained stable. The length of social pretend play did not change significantly over time. Children's age and language abilities were positively associated with successful initiations, failed initiations, and length of social pretend play. Very young children engage in social pretend play and the nature of this play evolves as they get to know one another and form relationships.

Meta-analysis is the synthesis of findings from research projects, which enables an estimate of the average or pooled effect across various studies. This study presents findings from the intention to treat analysis for a series of educational evaluations in England using a two-stage meta-analysis with standardised outcome data and individual participant data meta-analyses. The research estimates the overall impact of educational trials on pupils eligible for Free School Meals (FSM) and the attainment gap in literacy and mathematics performance between FSM and non-FSM pupils based on analysis of 88 trials and data from over half a million pupils. For the meta-analyses, frequentist and Bayesian multilevel models were used to estimate the individual and pooled effect size across categories of explanatory variables such as age groups (key stages in England) and aspects of the type of interventions (one-to-one, small group, whole class). Results indicated that the overall impact of interventions on the literacy outcomes of FSM pupils was positive, with a pooled effect size of 0.06 (0.03, 0.08). However, for mathematics, no overall effect on FSM pupils was observed. Analysis of the attainment gap indicated that literacy outcomes for FSM pupils were improved by interventions marginally more than for non-FSM pupils (pooled attainment gap 0.01 (-0.01, 0.04)). The risk of bias assessment showed that estimates were consistent across different methodological approaches. Overall, evidence from this study can be used to identify, test and scale educational interventions in schools to improve educational outcomes for disadvantaged pupils.

Improvement in student performance from educational reforms has driven current calls for changes in curricular, instructional, and assessment practices in higher education. Less research has addressed the possibility of such reforms aiding in resolving the well-documented inequities in educational outcomes between sexes and races/ethnicities in STEM courses. In this study, we compared a traditional and a reformed general chemistry course offered at a large public research-intensive university. Using hierarchical linear and logistical modeling, we assessed and investigated the inequities between sexes and races/ethnicities in the outcomes for these courses. Initial analysis looking at aggregated data (not by sex or ethnicity) showed the benefit of reforms specifically for students that had not been as academically well prepared as others. Novel findings for off-sequence (first semester taken in Spring) classes for both the traditional and reformed courses found a female-lead in course grades and benefits for Native Hawaiian or Pacific Islander, American Indian or Alaska Native, Latinx or Hispanic, and African American or Black (NALA) students. The on-sequence traditional course was found to have a male-lead in achievement in line with previous research. The on-sequence reformed course was found to have more equity in achievement between sexes. Possible sources of that improvement from the traditional course were investigated, and the weighting of different course components in the overall course grades was found to be a critical piece in the reformed course's more equitable outcomes. Though our results are promising in terms of offering a recommendation that may aid in advancing equity between sexes, our results indicate that the educational reforms made fell short in mitigating the pervasive racial/ethnic inequities found in achievement and performance measures.

This study examined the influences of teacher support and school environment on 10th graders' mathematics achievements, controlling for socioeconomic status, gender, and mathematics self-efficacy, and the differential influences of these factors for White and Black students in the USA. A sample of 8,682 White and 2,020 Black students who participated in the Educational Longitudinal Study were selected for the study. The results showed that Black students lagged behind their White peers on mathematics achievement, had lower socioeconomic status, and perceived the school environment as more negative. Hierarchical Linear Modeling analysis showed that the achievement gap between Black and White students was lower when students perceived a higher level of negative behavior in school.

School belonging is an important component of adolescent well-being, yet little is known about its relationship with adolescents' Information Communication Technology (ICT) use. This study aimed to examine the relationship between school belonging and various ICT use types in Australian adolescents. The sample was drawn from 14,530 Australian students in Grade 7 or higher, who completed the 2015 Organization for Economic Cooperation and Development's Program for International Student Assessment survey. A hierarchical regression analysis was conducted to investigate the relationship between self-reported measures of school belonging and ICT use at home for schoolwork and ICT use at home for leisure, adjusting for covariates (age, gender, and economic, social, and cultural status). The regression model accounted for 3% of the variability of sense of school belonging, R[superscript 2] = 0.03, F(5, 10196) = 60.00, p < 0.001. After adjusting for covariates, more frequent ICT use at home for schoolwork predicted a higher sense of school belonging. Conversely, more frequent ICT use at home for leisure predicted lower levels of sense of school belonging. The way adolescents engage with ICT is important for a student's sense of school belonging, and the present findings have implications for researchers and psychologists.

Oftentimes in many fields of the social and natural sciences, data are obtained within a nested structure (e.g., students within schools). To effectively analyze data with such a structure, multilevel models are frequently employed. The present study utilizes a Monte Carlo simulation to compare several novel multilevel classification algorithms across several varied data conditions for the purpose of prediction. Among these models, the panel neural network and Bayesian generalized mixed effects model (multilevel Bayes) consistently yielded the highest prediction accuracy in test data across nearly all data conditions.

Colleges and universities play a critical role in shaping intergroup dynamics in an era of increasing racial tensions in the United States. Diversity courses may serve as one important approach for preparing college students for participation in an equitable and just society, since this coursework holds a unique position at many institutions to expose college students to issues of difference and inequality. This study synthesizes research on the relationship between university/college instruction explicitly using the word course and the root divers and student outcomes over the span of 25 years. Within a meta-analytic sample of 355 effect sizes, from 73 publications, and 47 distinct samples representing 116,092 undergraduate students the results indicate an overall small positive association between diversity coursework and various outcomes. Additional results highlighted the ways in which this relationship is moderated by various characteristics of the courses, outcome measures, and study design.

This research examines the relationships between first-year college students' attendance at campus-sponsored extracurricular events and grade point average, persistence from fall to spring, and retention to the second year. Students' attendance at various campus life events over the course of an academic year was tracked by scanning students' identification cards. Hierarchical regression modeling revealed that first years attending events within a specific category of campus life programming called "signature events" had significantly higher cumulative grade point averages and better odds of persisting and retaining. The findings from this study add to the body of research on student involvement and engagement, with a specific focus on campus life activities.

The purpose of this study was to better understand the association between teachers' incoming classroom management skills and end-of-year literacy skills of preschool children with or at risk of emotional and behavioral disorders. Furthermore, we explored the contribution of students' incoming engagement and communication skills to end-of-year literacy skills. A series of multilevel models revealed that teacher classroom management predicted end-of-year letter sound fluency, but not letter naming fluency, after controlling for other factors. We conclude with a discussion of these preliminary findings and provide suggestions for future research and practice in early intervention settings.

Despite a growing body of turnover literature, much remains unknown about the factors predicting career transitional behaviors of school principals. To bridge this gap, we examined variations in principal, school, and district characteristics influencing administrator leaver and mover behaviors, using Hierarchical Generalized Linear Modeling. Our findings revealed that class size, support staff, parental involvement, teacher incentives, unionization, and many other district-level policies offset turnover, thereby contributing to retention. Furthermore, predictors of principal mover behavior differed from those of principal leaver behavior, suggesting that different forms of exit paths may need different policy tools to improve retention.

In this study we examined the relationship of students' 21st century skills with their academic and behavioral outcomes. We investigated K-12 teachers' (n = 150) judgment of students' (n = 3,108) use of 21st century skills (i.e., persistence, curiosity, affect, and cognition behaviors) via a Likert-type rating scale. We compared teachers' ratings with students' academic and behavioral outcomes in a southeastern school district. We used Hierarchical Linear Modeling with length of teacher acquaintance with the student as the nesting variable. Teacher ratings for students' low instances of persistence behaviors and high instances of externalizing affect behaviors were predictive of a higher probability of student office discipline referrals. We found a positive correlation for teachers' ratings of students' cognition behaviors and their reading, mathematics, and science outcomes. Teachers' student length of acquaintance was significant for academic outcomes. Results indicated a predictive relationship with teachers' judgment of students' 21st century skills and student outcomes.

The challenge and pleasures of studying child and adolescent peer experiences come from the complexity and the significance of these relationships for development in childhood and adolescence. In spite of the recognized strengths of the current literature on the effects of experiences with peers, research on peer experiences is often limited by an inattention to the effects of the hierarchical structure of the peer system. Experiences with peers are known to be situated in complex and multilevel social contexts. These contexts can be as small (or as focal) as a relationship dyad (e.g., two best friends interacting together), a peer group (e.g., more than two peers interacting together simultaneously), or as large and diverse as the cultural contexts of nation-states. A consequence of this multilevel organization peer research would benefit from the use of models that can account for variables and processes that occur at different levels of social complexity. This multilevel approach would also be multidisciplinary because it would require the integration of constructs typically studied in different scholarly disciplines. We show how a broader and less disciplinarily focused perspective can enrich our understanding of the features and processes of peer experiences.

Purpose: The purpose of this study was to determine the 3-year trajectory of students' interest in learning physical activity (PA) knowledge during middle school and the effects of gender and prior knowledge on the trajectory. Methods: Using a repeated-measures design, a cohort of 447 sixth-grade students' interest in learning PA knowledge was measured eight times from the beginning of sixth grade to the end of eighth grade. Students' prior PA knowledge was measured at the beginning of sixth grade. Results: Hierarchical linear modeling analyses showed that students' interest in PA knowledge in general declined over time. Girls' interest declined faster than that of the boys before eighth grade, but their decline slowed during the eighth-grade year. Prior knowledge did not influence the trajectories. Discussion: These findings provide insights regarding students' interest in learning PA knowledge and indicate that prior knowledge may not be a central component that contributes to interest development.

The purpose of this two-part study is to evaluate methods for multiple group analysis when the comparison group is at the within level with multilevel data, using a multilevel factor mixture model (ML FMM) and a multilevel multiple-indicators multiple-causes (ML MIMIC) model. The performance of these methods was evaluated integrally by a series of procedures testing weak and strong invariance models and the latent group mean differences testing after holding for factorial invariance. Two Monte Carlo simulation studies were conducted under the following conditions: number of clusters, cluster size, and the design type in groups. A multilevel one-factor confirmatory factor analysis (CFA) model as a research model in Study 1 was investigated to compare the results under different conditions with those of previous studies. A multilevel two-factor CFA model as a research model in Study 2 was evaluated by fitting alternative models that can be applied when the model is complicated. The results indicated that the two methods were reasonable in multilevel multiple groups analysis across within-level groups. However, pros and cons were found between the two methods. In the multilevel one-factor CFA model, ML MIMIC model was slightly better when the sample size is small. In the multilevel complex model, two alternative models of ML FMM were recommended because the weak invariance testing of ML MIMIC was considerably time-consuming. Finally, it was shown that information criteria, which are criteria for determining whether factorial invariance is established, need to be applied differently according to the sample size conditions. Guidelines for this situation are provided.

This study examined the impact of omitting covariates interaction effect on parameter estimates in multilevel multiple-indicator multiple-cause models as well as the sensitivity of fit indices to model misspecification when the between-level, within-level, or cross-level interaction effect was left out in the models. The parameter estimates produced in the correct and the misspecified models were compared under varying conditions of cluster number, cluster size, intraclass correlation, and the magnitude of the interaction effect in the population model. Results showed that the two main effects were overestimated by approximately half of the size of the interaction effect, and the between-level factor mean was underestimated. None of comparative fit index, Tucker-Lewis index, root mean square error of approximation, and standardized root mean square residual was sensitive to the omission of the interaction effect. The sensitivity of information criteria varied depending majorly on the magnitude of the omitted interaction, as well as the location of the interaction (i.e., at the between level, within level, or cross level). Implications and recommendations based on the findings were discussed.

The purpose of this study was to examine which Ohio schools offered curricular music courses and the rates at which students participated in those courses. The analysis involved descriptive statistics, analysis of variance, logistic regression, and partially nested multilevel modeling using data from the Ohio Department of Education (N = 3,222 schools). The investigation revealed that charter schools offered music courses far less often than public schools. However, in charter schools that did offer music, students participated at higher rates than those in public schools. Nearly all public schools featured music classes. The exception was high schools in the highest poverty urban neighborhoods, 31% of which had no curricular music. Students identified as Black, Hispanic, or indigenous were more likely to attend schools without music programs. Elementary students enrolled in an average of 1.00 music classes per year, whereas middle and high school students enrolled in 0.67 and 0.35 music classes per year, respectively. Suburban districts saw the greatest decline in music participation as students progressed to high school. Urban schools with greater percentages of white, non-Hispanic students had higher music enrollment rates.

Background: The retention of traditionally underserved students remains a pressing problem across graduate engineering programs. Disciplinary differences in graduate engineering identity provide a lens to investigate students' experiences and can pinpoint potential opportunity structures that support or hinder progress based on social and personal identities. Purpose: This study investigates the impact of discipline, gender, race/ethnicity, advisor relationship, and years in a program on graduate engineering identity variability. Methods: Cross-sectional survey data from a national sample of doctoral engineering students were analyzed with multilevel modeling. Multilevel modeling measured the differences at the individual and discipline levels for graduate engineering identity and the domains of engineer, researcher, and scientist. Independent variables included were gender, advisor relationship score, race/ethnicity, and years in a program. Results: The engineer identity sub-construct of recognition significantly varied among engineering disciplines. Traditionally underserved students (i.e., Women and minoritized racial/ethnic groups) expressed lower engineering recognition levels, with this relationship varying based on discipline. Overall, our model explained 30% of the variation in engineering recognition among disciplines. Conclusions: The disciplinary variation in graduate engineering identity combined with the significance of gender and race/ethnicity indicates traditionally underserved students do not experience equivalent opportunity structures compared with their well-represented peers. Modifying traditional opportunity structures to serve students better may provide the needed changes to engage and retain traditionally underserved populations.

Creativity and well-being are popular subjects in psychological and organizational studies. The recent literature presented mixed perspectives about the nature of the relationship between the two. Whereas the mad-genius hypothesis, which was often explored among eminently creative individuals, seems to imply a negative relationship between the two, trends in the field of creativity, such as everyday creativity and general psychology (i.e., positive psychology) linked them positively. The present meta-analysis study synthesized 189 effect sizes obtained from 32 samples in 26 different studies based on a total sample of 8,189. Analyses with multilevel modeling yielded a mean effect of r = 0.14. The moderator analysis tested the impact of age, gender, measure of creativity, measure of well-being, stimulus type of creativity measure, and index of creativity measure. Only the creativity measure explained the variation in the study outcomes. The relationship between creativity and well-being was significantly higher when creativity was measured by instruments focusing on creative activity and behavior (r = 0.22) than the divergent thinking tasks (r = 0.06). Those findings are discussed in terms of theoretical and practical implications.

In an era of social media, and amid the coronavirus pandemic, teachers' curation of instructional resources stands as an opportunity to observe teachers' planning and conceptualization of their practice in real time. This work explores the resources teachers curate, their rigor, and the effects on students' learning across years. Merging big data from Pinterest, a prominent social media platform, and administrative data from the Indiana Department of Education on 10,383 fourth- through sixth-grade students across 2010-2017, this study employed three-level hierarchical linear models to identify relationships between the inherent cognitive rigor in teachers' curated instructional tasks and students' achievement. Results indicate teachers curating resources focused on basic memorization and remembering negatively affected students' learning, whereas higher-level tasks focused on understanding and applying had a positive impact on achievement. Identified effects were comparable to those related to student and teacher attributes, signaling the importance of teachers' curation.

Prior studies have focused on general technology use and technology use in domain-general applications and quantity of technology use. Recent evidence suggests that investigations should consider how technology is used in more contextually specific ways, including how technology is used for various cognitive tasks in specific classrooms. The purpose of this study was to examine the ways in which classroom content area and student goal orientation have a coordinated influence for how students used technology to support learning. The sample included high school students in a Midwestern state who were surveyed on their motivation and how they used technology to support learning. The study employed hierarchical linear modelling to examine how goal orientation and classroom content area predicted various levels of Bloom's Digital Taxonomy. Students who adopted mastery-oriented goals were more likely to use technology for various cognitive tasks, especially those at higher levels of complexity. Lastly, the association between mastery goal orientation and some aspects of technology use was conditioned on content area, although effect sizes were small. This study showed that, overall, technology is used differentially across four core content areas. Students in mathematics classrooms used technology less, however much of technology use was evident at lower cognitive levels. Second, students' goal orientation, and in particular their mastery goals influence how technology is used across content areas, and this is marginally conditioned on content area. Technology use should match the instructional context to maximize technology use and students' goal orientation.

Since the early 2000s, educational evaluation research has primarily centered on teachers', rather than schools', contributions to students' academic outcomes due to concerns that estimates of the latter were smaller, less stable, and more prone to measurement error. We argue that this disparity should be reduced. Using administrative data from three cohorts of Massachusetts public school students (N = 123,261) and two-level models, we estimate middle schools' value-added effects on eighth-grade and 10th-grade math scores and, importantly, a non-test score outcome: 4-year college enrollment. Comparing our results to teacher-centered studies, we find that school effects (encompassing both teaching- and nonteaching-related factors) are initially smaller but nearly as stable and perhaps more persistent than are individual teacher effects. Our study motivates future research estimating the long-term effects of both teachers and schools on a wide range of outcomes.

Visual diagnosis of radiographs, histology and electrocardiograms lends itself to deliberate practice, facilitated by large online banks of cases. Which cases to supply to which learners in which order is still to be worked out, with there being considerable potential for adapting the learning. Advances in statistical modeling, based on an accumulating learning curve, offer methods for more effectively pairing learners with cases of known calibrations. Using demonstration radiograph and electrocardiogram datasets, the advantages of moving from traditional regression to multilevel methods for modeling growth in ability or performance are demonstrated, with a final step of integrating case-level item-response information based on diagnostic grouping. This produces more precise individual-level estimates that can eventually support learner adaptive case selection. The progressive increase in model sophistication is not simply statistical but rather brings the models into alignment with core learning principles including the importance of taking into account individual differences in baseline skill and learning rate as well as the differential interaction with cases of varying diagnosis and difficulty. The developed approach can thus give researchers and educators a better basis on which to anticipate learners' pathways and individually adapt their future learning.

The concurrent associations between students' perceptions of cognitive-behavioral and emotional engagement in schools and three factors aligning with the major aims of the school-wide social-emotional learning (SEL) approach (i.e., teacher-student relationships, student-student relationships, and teaching of social and emotional competencies) were examined among 25,896 students across elementary, middle, and high school while controlling statistically for demographic variables. Results indicated that at the student level all three factors were associated significantly with cognitive-behavioral engagement, but at the school level only the teaching of social and emotional competencies was associated significantly with cognitive-behavioral engagement. All three factors were also associated significantly with emotional engagement at both the student and school levels, with teacher-student relationships having the strongest association. Results of moderating analyses revealed that the strength of association of student engagement with teacher-student relationships, student-student relationships, and the teaching of social-emotional competencies varied depending on the types of engagement and students' grade levels. These and other key findings, as well as implications for research and practice, are discussed.

Latent class analysis was used to examine whether patterns of bystander responses varied as a function of both student- and school-level characteristics. Data from 18,863 high school students from 58 schools who "ever witnessed bullying" were used to identify five latent classes of bystander behavior. Three of the classes identified paralleled commonly used researcher-identified categories ("Passive" = 9.7%, "Defender" = 20.4%, and "Contributor" = 3.4%), whereas we also identified two patterns of bystander responses that had not been previously characterized ("Limited" = 64.8% and "Inconsistent" = 1.7%). Multilevel logistic regression models were then used to examine student- and school-level characteristics that differentiate those in the defender class from other bystander classes. Youth in the defender class were more likely to believe that other students intervene with bullying, and they felt a greater connection with school staff as compared to youth in all other bystander classes. Further, gender, normative beliefs about retaliation, and bullying involvement were associated with class membership. Findings indicated that defending bystander responses are relatively low and suggested that school-level contextual factors, youth perceptions of others' bystander behavior, and bullying involvement all inform our understanding of adolescent bystander behavior.

Learning analysts often consider whether learning processes across time are related (1) to one another or (2) to learning outcomes at higher levels. For example, are a group's temporal sequences of talk (e.g., correct evaluation [right arrow] correct, new idea) during its problem solving related to its group solution? I show how to address these issues with (1) a higher-level outcome regression and (2) a lower-level process regression, applying both to 3,234 turns of talk by 80 students working in 20 groups to solve an algebra problem. The easy-to-use, outcome-level analysis of group solution score has the following problems: multicollinearity, possibly low statistical power, cannot test for links among sequence components, and cannot model outcomes at multiple levels. The complex, process-level analysis for turns of talk overcomes these shortcomings with multilevel analysis, vector auto-regression, and outcome-level regression residuals. These results suggest a combined procedure. First, run an outcome-level analysis. If the results are significant, then the outcome-level results suffice. Otherwise, non-significant results might reflect multicollinearity, which then requires a process-level analysis. This procedure can help test a comprehensive model of how learning processes or their temporal sequences are related to learning outcomes at the turn-, time period-, individual-, group-, class-, and school-levels.

Fostering peer interaction and shared learning is an important aim of inclusive instruction. However, it has not been established whether it is possible to offer explicit and intensive support for low achievers in inclusive settings. This longitudinal study examined whether a structured program that includes cooperative learning fosters computational competence and flexible strategy use by low achievers in mathematics in inclusive classrooms. 126 persistent counters from 35 inclusive classrooms in grade 2 participated in a structured intervention lasting ten weeks, under three conditions: Cooperative learning (CL, students working in pairs during seatwork), individual learning (IND, students working individually during seatwork) and a control group (CG) with "business as usual." Even when there was substantial class-specific differential development during group-based interventions, between pre- and posttest, results showed no statistically significant main effects for either of the two intervention conditions relative to the control group. However, compared to the CG-group, the CL intervention showed a higher slope for pretest performance, indicating that students with a higher level of computation competence benefited relatively more from the CL intervention than students with similar preconditions in the CG. The results provide evidence that multiple approaches are needed; approaches which combine well-structured programs stimulating cooperative learning and shared learning situations with intensive and individualized measures are of benefit for students with very low mathematical competence.

Computers, during its entry into the academic domain, were forecasted to revolutionize the entire educational system across the globe. At present, similar expectations have been encircling the social media technologies. Many researchers claim that social media, when introduced in higher education environments, tend to enhance the students' learning experience by facilitating heightened levels of instructor-student interactions, peer learner support and learners' collaboration. Though social media is claimed to take the higher education learning to the next level, it is crucial to identify the readiness and proficiency of the instructors who play the main role in efficiently implementing these technologies in their classrooms. Especially, instructors belonging to the institutions present in developing nations are less likely to have proper access to some of the basic facilities (such as power supply or internet connectivity) required for social media integration. Hence, the present study aims at identifying the feasibility of introducing social media aids to the higher education classrooms present in Tamil Nadu--one among the eight Indian states having the highest number of colleges in the Nation.