Multilevel Latent Growth Curve Models Research Articles

Self-regulated and externally regulated learning in adolescence: Developmental trajectories and relations with teacher behavior, parent behavior, and academic achievement.

Both self-regulation and external regulation are key to understanding adolescents' learning and positive development at school. However, evidence on the joint development of self-regulated learning and externally regulated learning during adolescence is lacking. In addition, the current knowledge on interrelations between the development of adolescents' self-regulated learning, externally regulated learning, behaviors of teachers and parents in terms of autonomy support and achievement pressure, and academic achievement is very limited. The present multilevel longitudinal analysis focusing on the domain of mathematics (N = 1,542 German adolescents; annual assessments from Grades 5 to 9; mean age at Grade 5 = 11.79 years, SD = 0.71, 51.75% female) addressed these gaps. Results from multilevel latent basic growth curve models showed that self- and externally regulated learning decreased over the 5 years at both the individual student and the class level. Changes in self- and externally regulated learning were linked: Classes with higher levels of self-regulated learning at Grade 5 showed a stronger decrease in externally regulated learning over time. Initial levels of and changes in student-reported teacher and parental autonomy support and achievement pressure were associated with self- and externally regulated learning at the individual student level; student-reported teacher autonomy support and self-regulated learning were also linked at the class level. Self-regulated learning related positively to standardized achievement test scores but not to adolescents' grades. This study adds to the scarce evidence base on different regulatory forms of adolescents' learning and can inform future research on adolescents' positive development and educational practice. (PsycInfo Database Record (c) 2023 APA, all rights reserved).

Longitudinal changes in Chinese minority college students’ health-related fitness: A multilevel latent growth curve modeling approach

Purpose: The study aimed to test the overall changes of health-related fitness (HRF) in minority Chinese college students and to examine HRF differences in gender, race, and year in college. Method: Participants (n = 1320) were minority college students with more than two-thirds females (ie 76.1%), and Hui, Tibetan, and Mongolia consisted of 13.8%, 13.8%, and 11.2%, respectively. Student HRF was tracked for four years. Data were analyzed using multilevel latent growth curve modeling. Results: Muscular strength and endurance were the weakest component in minority college students’ HRF, while body mass index was within the category of “excellent”. Males outperformed female on all components of HRF. Conclusions: It is suggested that interventions concerning minority females’ HRF and muscular strength and endurance for both genders be constructed and tested.

Examining the effect of L2 motivational factors on the development of L2 achievement: using multilevel latent growth curve model

Examining the effect of L2 motivational factors on the development of L2 achievement: using multilevel latent growth curve model

Family socioeconomic status and Chinese preschoolers' social skills: Examining underlying family processes.

Using a longitudinal sample of 508 Chinese preschool children, this study examined how family socioeconomic status (SES) was related to preschool children's social skills development through family processes using multilevel latent growth curve modeling. After controlling for the effects of personal characteristics, a significant indirect effect of family SES on initial levels of children's social skills and growth was observed and mediated through maternal depressive symptoms, marital relationships, and parenting practices. Maternal depressive symptoms from family SES was linked to poorer marital relationships and parenting practices, which were linked to children's social skills. Authoritative parenting was related to increased growth in children's social skills. In addition, authoritative parenting mediated the effects of marital relationship quality on both initial levels and growth in social skills. Implications for future research and intervention are discussed. (PsycInfo Database Record (c) 2020 APA, all rights reserved).

Emotional Responses to Stressors in Everyday Life Predict Long-Term Trajectories of Depressive Symptoms.

Individuals' emotional responses to stressors in everyday life are associated with long-term physical and mental health. Among many possible risk factors, the stressor-related emotional responses may play an important role in future development of depressive symptoms. The current study examined how individuals' positive and negative emotional responses to everyday stressors predicted their subsequent changes in depressive symptoms over 18 months. Using an ecological momentary assessment approach, participants (n = 176) reported stressor exposure, positive affect (PA), and negative affect (NA) five times a day for 1 week (n = 5,483 observations) and provided longitudinal reports of depressive symptoms over the subsequent 18 months. A multivariate multilevel latent growth curve model was used to directly link the fluctuations in emotions in response to momentary stressors in everyday life with the long-term trajectory of depressive symptoms. Adults who demonstrated a greater difference in stressor-related PA (i.e., relatively lower PA on stressor vs. nonstressor moments) reported larger increases in depressive symptoms over 18 months. Those with greater NA responses to everyday stressors (i.e., relatively higher NA on stressor vs. nonstressor moments), however, did not exhibit differential long-term changes in depressive symptoms. Adults showed a pattern consistent with both PA and NA responses to stressors in everyday life, but only the stressor-related changes in PA (but not in NA) predicted the growth of depressive symptoms over time. These findings highlight the important-but often overlooked-role of positive emotional responses to everyday stressors in long-term mental health.

Acceleration of drinking pace throughout the evening among frequently drinking young adults in the Netherlands

Background and AimsPrevious studies have found that in some countries ‘drinking pace’ (number of drinks consumed per hour) increases during the course of an evening. We aimed to provide evidence of this acceleration from a culture in which binge drinking is prevalent and to test whether this is consistent across gender, day of week and in high‐risk drinkers.DesignEvent‐level data collected on Thursday, Friday and Saturday evenings over 5 consecutive weeks.SettingThe Netherlands.ParticipantsA total of 197 young adult frequent drinkers (48.7% women, mean age = 20.8).MeasurementsHigh‐risk drinking (assessed by the Alcohol Use Disorders Identification Test) and gender were measured at baseline, and questionnaires were sent to participants’ smartphones every hour between 9 p.m. and 1 a.m. A total of 7185 questionnaires during 1589 evenings were used for the analyses.FindingsMulti‐level latent growth curve models revealed an acceleration in drinking on days of the week tested [throughout all evenings; b = 0.430, standard error (SE) = 0.045, P < 0.001], which stabilized as the evening progressed (b = −0.072, SE = 0.008, P < 0.001). The temporal pattern did not differ between the days or gender, but men started with a higher number of drinks at the beginning of the evening (b = 0.465, SE = 0.099, P < 0.001). High‐risk drinking was related to more alcoholic drinks at the beginning of an evening (b = 0.032, SE = 0.011, P = 0.003) and a steeper acceleration during the subsequent hours (b = 0.021, SE = 0.009, P = 0.024).ConclusionsYoung adults in the Netherlands appear to show an increase in drinking pace during the course of an evening's drinking, with high‐risk drinkers showing a greater increase.

Evaluating fit indices in a multilevel latent growth curve model: A Monte Carlo study.

The multilevel latent growth curve model (MLGCM), which is subsumed by the multilevel structural equation modeling framework, has been advocated as a means of investigating individual and cluster trajectories. Still, how to evaluate the goodness of fit of MLGCMs has not been well addressed. The purpose of this study was to conduct a systematic Monte Carlo simulation to carefully investigate the effectiveness of (a) level-specific fit indices and (b) target-specific fit indices in an MLGCM, in terms of their independence from the sample size's influence and their sensitivity to misspecification in the MLGCM that occurs in either the between-covariance, between-mean, or within-covariance structure. The design factors included the number of clusters, the cluster size, and the model specification. We used Mplus 7.4 to generate simulated replications and estimate each of the models. We appropriately controlled the severity of misspecification when we generated the simulated replications. The simulation results suggested that applying RMSEAT_S_COV, TLIT _ S _ COV, and SRMRB maximizes the capacity to detect misspecifications in the between-covariance structure. Moreover, the use of RMSEAPS _ B, CFIPS _ B, and TLIPS _ B is recommended for evaluating the fit of the between-mean structure. Finally, we found that evaluation of the within-covariance structure turned out to be unexpectedly challenging, because none of the within-level-specific fit indices (RMSEAPS _ W, CFIPS _ W, TLIPS _ W, and SRMRW) had a practically significant sensitivity.

Do Demographic Changes Jeopardize Social Integration among Aging Adults Living in Rural Regions?

ObjectivesContextual influences of the living environment on the social integration of adults have been primarily studied cross-sectionally. Here, we argue that context (i.e., urban vs rural environment) as well as changes in context (i.e., population decline) are potentially important for the development of social integration across middle adulthood and late life.MethodWe used a large-scale longitudinal data set representative of the late middle-aged and older German population (N = 4,790; aged 40 to 85 years) that assessed participants every 6 years across 3 waves. To examine our assumptions, we implemented multilevel latent growth curve models.ResultsWe found that declines in network size were more pronounced in rural than in urban areas. Moreover, age-related declines in network size, social engagement, and social support were particularly pronounced in rural districts that demonstrated above average population decline.DiscussionOur results imply that ongoing demographic changes, particularly in rural areas, may introduce additional barriers for maintaining social integration into late life.

Analyzing individual growth with clustered longitudinal data: A comparison between model-based and design-based multilevel approaches.

To prevent biased estimates of intraindividual growth and interindividual variability when working with clustered longitudinal data (e.g., repeated measures nested within students; students nested within schools), individual dependency should be considered. A Monte Carlo study was conducted to examine to what extent two model-based approaches (multilevel latent growth curve model - MLGCM, and maximum model - MM) and one design-based approach (design-based latent growth curve model - D-LGCM) could produce unbiased and efficient parameter estimates of intraindividual growth and interindividual variability given clustered longitudinal data. The solutions of a single-level latent growth curve model (SLGCM) were also provided to demonstrate the consequences of ignoring individual dependency. Design factors considered in the present simulation study were as follows: number of clusters (NC = 10, 30, 50, 100, 150, 200, and 500) and cluster size (CS = 5, 10, and 20). According to our results, when intraindividual growth is of interest, researchers are free to implement MLGCM, MM, or D-LGCM. With regard to interindividual variability, MLGCM and MM were capable of producing accurate parameter estimates and SEs. However, when D-LGCM and SLGCM were applied, parameter estimates of interindividual variability were not comprised exclusively of the variability in individual (e.g., students) growth but instead were the combined variability of individual and cluster (e.g., school) growth, which cannot be interpreted. The take-home message is that D-LGCM does not qualify as an alternative approach to analyzing clustered longitudinal data if interindividual variability is of interest.

Increasing ethnic diversity moderates longitudinal effects of individual differences on friendship homophily

AbstractThis study examined direct and interactive effects of social–emotional adjustment, national and ethnic identification, and school ethnic composition on friendship homophily among 214 ethnic minority and 183 ethnic majority English children, aged between 5 and 11 years. The data came from a longitudinal study, which included 3 time points, spanning a 12‐month period. Results of multilevel latent growth curve models showed that among ethnic minority English children (teacher‐rated), peer problems and ethnic identity were associated with more friendship homophily, whereas a bicultural identity was not related to more friendship homophily. Among ethnic majority English children, the effects of peer problems and English identity were moderated by school ethnic composition, such that these factors were not associated with more friendship homophily in more ethnically diverse schools. The findings are discussed on the basis of theories of intergroup contact and intergroup threat.

Consequences of Misspecifying Across-Cluster Time-Specific Residuals in Multilevel Latent Growth Curve Models

This Monte Carlo study evaluates, in the context of multilevel latent growth curve models, the consequences of under- and overspecifying across-cluster time-specific residuals (i.e., ) on the estimation of the fixed effects, their corresponding standard errors, the variances and covariances of the random effects, Type I error rates, and the statistical power of detecting fixed effects. The results show that underspecifying with all elements of fixed at zero results in a large underestimation of the between- and within-level random effect and standard errors of fixed effect estimates, which, in turn, leads to serious bias in significant testing. Underspecifying with diagonal elements of constrained to equality, or overspecifying with diagonal elements of constrained to equality or freely estimated and residual covariances fixed at zero also leads to bias in the estimation of the between- and within-level random effects. Implications of the compensatory relationship occurring at the covariance level are discussed.

Trajectories of Change in Obesity among Tehranian Families: Multilevel Latent Growth Curve Modeling.

Objectives. To evaluate the trajectories of change in obesity within and between Tehranian families, who participated in the Tehran Lipid and Glucose Study (TLGS). Methods. This study is a family-based longitudinal design, in four waves. A total of 14761 individuals, within 3980 families, were selected. Three anthropometric measurements, body mass index (BMI), waist circumference (WC), and a body shape index (ABSI), were recorded. Multilevel latent growth curve modeling (MLGCM) approach was used for evaluating the change trajectories in obesity within and between the families. Results. The mean age of the subjects in the present study was 33.28 ± 19.01 (range 3–89 years) and 50.1% were male. Obesity was significantly increased (P < 0.001). Individuals with more fat become obese slower, whereas families with more fat become obese faster (P < 0.001). The initial value and growth rate of WC and ABSI were greater in men than in women, while this result is contrary to BMI (P < 0.001). Conclusions. Our findings demonstrated that there is an alarming increase in the obesity trend in Tehranian families. The important role of the family in the prevention of obesity is highlighted, underlining the need for public health programs, as family centered educations to lifestyle modification, which can address this emerging crisis.

Relative Contributions of a Set of Health Factors to Selected Health Outcomes.

Although many researchers agree that multiple determinants impact health, there is no consensus regarding the magnitude of the relative contributions of individual health factors to health outcomes. This study presents a method to empirically estimate the relative contributions of health behaviors, clinical care, social and economic factors, and the physical environment to health outcomes using nationally representative county-level data and statistical approaches that account for potential sources of bias. The analyses for this study were conducted in 2014. Data were from the 2010-2013 County Health Rankings & Roadmaps. Data covered 2,996 of 3,141 U.S. counties. Ordinary least squares modeling was used as a baseline model. Multilevel latent growth curve modeling was used to estimate the relative contributions of health factors to health outcomes while accounting for measurement errors and state-specific characteristics. Almost half of the variance of health outcomes was due to state-level variation rather than county-level variation. When adjusted for measurement errors and state-level variation using multilevel latent growth curve modeling, the relative contribution of clinical care decreased and that of social and economic factors increased compared with the baseline model. This study presents how potential sources of bias affected the estimates of the relative contributions of a set of modifiable health factors to health outcomes at the county level. Further verification of these approaches with other data sources could lead to a better understanding of the impact of specific health determinants to health outcomes, and will provide useful information on policy interventions.

Donor catch-up growth after laser surgery for twin–twin transfusion syndrome

ObjectiveTo assess fetal growth after laser surgery for TTTS at the time of prenatal diagnosis, birth, and at 2years of age. Design/methodsGrowth data were collected from surviving children treated between 2007 and 2010 as part of a study to assess neurodevelopment at 24months (±6weeks) corrected age. Fetal weights were obtained via ultrasound using Hadlock's formula at the time of preoperative assessment for laser surgery. Birth weights were recorded by the staff at the delivering institutions. Weights at 2years corrected age were recorded at the time of neurodevelopmental testing. Weights were converted into percentiles according to standard growth curves. Growth restriction was defined as <10th percentile for given age. Multilevel latent growth curve models in Mplus (twins nested in families) examined weight change over time as a function of donor status, and repeated measures ANOVA was utilized to assess in donor–recipient weight discordance over time for twin pairs. Results99 of 206 children (56 of 130 families) were studied. There were no differences between enrolled and non-enrolled patients in donor/recipient status and survival rates, fetal demise, intrauterine growth restriction, Quintero stage, and gestational age of surgery or delivery. 48.5% were donors. The median fetal, birth, and 2-year weights for all twins were 288g, 1.9kg, and 11.8kg, respectively, and the overall prevalence of growth restriction was 28%, 22%, and 3%, respectively. Growth restriction rates at prenatal diagnosis were 56% in donors vs. 2% in recipients (OR=64.3, p<0.001); at birth, 35% vs. 10% (OR=5.0, p<0.01); and at 2years, 6% vs. 0%. Donors showed significant gains in weight percentile (B=13.1, p<0.001) and a significant decrease in growth restriction rates over time (B=−1.6, p<0.001). Weight discordance between donor and recipient pairs also significantly decreased over time (linear F(1,42)=54.34, p<0.001). ConclusionsAfter laser surgery for TTTS, donor twins exhibit significant catch-up growth by two years of age.

Using Design-Based Latent Growth Curve Modeling With Cluster-Level Predictor to Address Dependency

The authors compared the effects of using the true Multilevel Latent Growth Curve Model (MLGCM) with single-level regular and design-based Latent Growth Curve Models (LGCM) with or without the higher-level predictor on various criterion variables for multilevel longitudinal data. They found that random effect estimates were biased when the higher-level predictor was not included and that standard errors of the regression coefficients from the higher-level were underestimated when a regular LGCM was used. Nevertheless, random effect estimates, regression coefficients, and standard error estimates were consistent with those from the true MLGCM when the design-based LGCM included the higher-level predictor. They discussed implication for the study with empirical data illustration.

Parent Involvement and Science Achievement: A Cross-Classified Multilevel Latent Growth Curve Analysis

ABSTRACTThe authors examined science achievement growth at Grades 3, 5, and 8 and parent school involvement at the same time points using the Early Childhood Longitudinal Study–Kindergarten Class of 1998–1999. Data were analyzed using cross-classified multilevel latent growth curve modeling with time invariant and varying covariates. School-based involvement did not predict students’ initial science achievement status or their growth trajectory. Gender, race/ethnicity, parent's socioeconomic status, and school's percentage of free lunch eligible students were better predictors of intercept or slope. Findings indicated that students with lower initial science achievement scores had a faster rate of growth across time. Therefore, the science achievement gap narrows over time.

Longitudinal changes in physical fitness performance in youth

Using a multilevel latent growth curve modeling (LGCM) approach, this study examined longitudinal change in levels of physical fitness performance over time (i.e. four years) in young adolescents aged from 12–13 years. The sample consisted of 6622 students from 138 secondary schools in Singapore. Initial analyses found between-school variation on fitness test scores with intra-class correlations ranging from 0.02–0.19. Subsequent multilevel growth curve analyses revealed a quadratic trend of the longitudinal data across five stations of performance tests, with significant within-school student variability in change over time. The result of the multilevel LGCM showed that there were strong school effects on all of the physical fitness performances, in addition to inter- and intra-individual differences.

Individual, Social, and Neighborhood Influences on the Launch of Adolescent Antisocial Behavior

This study attempts to develop further insight into the initiation and continuance of antisocial behavior during adolescence using a “launch” perspective in conjunction with several risk/protective factor measures from prior research on substance use and delinquency. Data from the Project on Human Development in Chicago Neighborhoods (PHDCN) were analyzed using multilevel latent growth curve models across three waves for three age cohorts ( n = 2,130 youth; 78 neighborhoods). After considering typical trends and their variation, the relationships between covariates and different characteristics of the growth curves were analyzed. Implications of the findings for understanding developmental trends and prevention strategy are considered.

School effects on the development of motivation toward learning tasks and the development of academic self-concept in secondary education: a multivariate latent growth curve approach

The main objective of the present study is to gain more insight into the school effects on the development of 2 noncognitive student outcomes, namely, the motivation toward learning tasks and the academic self-concept, and, more specifically, on the consistency of the school effects on these 2 outcomes. Data were drawn from the “Longitudinaal Onderzoek Secundair Onderwijs” (LOSO)-project and consisted of 2,687 students in 50 schools that were tested repeatedly at 4 different time points during secondary education (Grades 7–12). A multivariate multilevel latent growth curve model was used to analyze the data. The results showed that the school effects on the development of the noncognitive outcomes were considerable, and, more importantly, the school effects were larger on growth than on student status. Schools that were effective for the development of the motivation toward learning tasks also proved to be effective for the development of academic self-concept. However, the consistency of the school effects resulted largely from intake differences between schools.

Differential school effects among low, middle, and high social class composition schools: a multiple group, multilevel latent growth curve analysis

This study uses large-scale survey data and a multiple group, multilevel latent growth curve model to examine differential school effects between low, middle, and high social class composition public schools. The results show that the effects of school inputs and school practices on learning differ across the 3 subpopulations. Moreover, student learning in low social class schools is far more sensitive to school factors than in middle and high social class schools. Yet, even after controlling for an extensive set of student background characteristics and school inputs, students attending low social class schools continued to learn at significantly slower rates. Widespread differences in the characteristics of schools across the subpopulations, which consistently challenge the educational milieu in low social class schools, likely contribute to differential school effects, as well as to the disparity in learning rates. The findings of this study accentuate the importance of school effectiveness research that goes beyond generic models to examine differential effects.