Putting the individual back into individual growth curves

Self-rated health and depressive symptoms in older adults: A growth mixture modeling approach

OBJECTIVE

Self-Rated Health (SRH) and depressive symptoms are important indicators of global quality of life in older adults. Prior research suggests associations between SRH and depressive symptoms. The current study assessed latent groups in levels and trajectories of these two subjective health indicators and how the latent groups relate to each other.

METHODS

Participants from the Australian Longitudinal Study of Aging (N = 2,087, ages 65+) were assessed over six waves of data collection, spanning eight years.

RESULTS

Growth Mixture Models were run for SRH and depressive symptoms, each yielded three latent groups with similar patterns: for both SRH and depressive symptoms two groups differing in their level with worsening status over time, and a third stable, but poorer functioning group. Analysis of the assignment of the latent groups revealed a consistent pattern for the majority, but some people were high in depression and high in SRH and some were low in depression and low in SRH.

CONCLUSIONS

SRH and depressive symptoms yielded both three latent groups whose combination supported the expected assignment for the majority and an unexpected assignment for some people. This may be a result of a protective factor existing for one variable but not the other.

Oxytocin receptor gene (OXTR) variant rs1042778 moderates the influence of family environment on changes in perceived social support over time

BACKGROUND

Lack of social support is an established risk factor across health outcomes, making it important to examine its family environmental and genetic determinants.

METHODS

In a 27-year follow-up of the Young Finns Study (N = 2341), we examined with a latent growth curve model whether genes involved in the oxytocin signaling pathway-namely, oxytocin receptor gene (OXTR) variants rs1042778, rs2254298, and rs53576-moderate the effect of early-life social experiences on perceived social support across the life span. Mothers reported the emotional warmth and acceptance towards their children at baseline when the participants were from 3 to 18 years old (1980). Perceived family support and support from friends and peripheral sources were assessed in five follow-ups 18 years apart (1989-2007).

RESULTS

Maternal emotional warmth and acceptance predicted the initial level of perceived social support across subscales, while the rate of change in family support was affected by the family environment only if participants carried the T-allele of OXTR rs1042778. This gene-environment interaction was not found for the rate of change in support from friends and peripheral sources and we also did not find associations between latent growth in perceived social support and OXTR variants rs53576 and rs2254298.

LIMITATIONS

Selective attrition in perceived social support, maternal emotional warmth and acceptance, gender, and SES. Family environment was assessed by a non-standardized measure.

CONCLUSIONS

OXTR rs1042778 polymorphism seems to contribute to changes in perceived family support in that way that some individuals (T-allele carriers) 'recover', to some extent, from the effects of early-life social experiences, whereas others (G/G genotype carriers) do not.

Virtual Levels and Role Models: N-Level Structural Equations Model of Reciprocal Ratings Data

A general latent variable modeling framework called n-Level Structural Equations Modeling (NL-SEM) for dependent data-structures is introduced. NL-SEM is applicable to a wide range of complex multilevel data-structures (e.g., cross-classified, switching membership, etc.). Reciprocal dyadic ratings obtained in round-robin design involve complex set of dependencies that cannot be modeled within Multilevel Modeling (MLM) or Structural Equations Modeling (SEM) frameworks. The Social Relations Model (SRM) for round robin data is used as an example to illustrate key aspects of the NL-SEM framework. NL-SEM introduces novel constructs such as 'virtual levels' that allows a natural specification of latent variable SRMs. An empirical application of an explanatory SRM for personality using xxM, a software package implementing NL-SEM is presented. Results show that person perceptions are an integral aspect of personality. Methodological implications of NL-SEM for the analyses of an emerging class of contextual- and relational-SEMs are discussed.

The Consequences of Varying Measurement Occasions in Discrete-Time Survival Analysis

Abstract. In a discrete-time survival model the occurrence of some event is measured by the end of each time interval. In practice it is not always possible to measure all subjects at the same point in time. In this study the consequences of varying measurement occasions are investigated by means of a simulation study and the analysis of data from an empirical study. The results of the simulation study suggest that the effects of varying measurement occasions are negligible, at least for the scenarios that were covered in the simulation. The empirical example shows varying measurement occasions have minor effects on parameter estimates, standard errors, and significance levels.

Multivariate normal maximum likelihood with both ordinal and continuous variables, and data missing at random

A novel method for the maximum likelihood estimation of structural equation models (SEM) with both ordinal and continuous indicators is introduced using a flexible multivariate probit model for the ordinal indicators. A full information approach ensures unbiased estimates for data missing at random. Exceeding the capability of prior methods, up to 13 ordinal variables can be included before integration time increases beyond 1 s per row. The method relies on the axiom of conditional probability to split apart the distribution of continuous and ordinal variables. Due to the symmetry of the axiom, two similar methods are available. A simulation study provides evidence that the two similar approaches offer equal accuracy. A further simulation is used to develop a heuristic to automatically select the most computationally efficient approach. Joint ordinal continuous SEM is implemented in OpenMx, free and open-source software.

Developmental associations between bilingual experience and inhibitory control trajectories in Head Start children

Children from lower socioeconomic (SES) backgrounds tend to be at-risk for executive function (EF) impairments by the time they are in preschool, placing them at an early disadvantage for academic success. The present study examined the potentially protective role of bilingual experience on the development of inhibitory control (IC) in 1146 Head Start preschoolers who were followed for an 18-month period during the transition to kindergarten as part of the longitudinal Family and Child Experiences Survey (FACES) 2009 study. Using three waves of data, we predicted individual variation in developmental trajectories of IC for three groups that differed in bilingual experience-English monolinguals, Spanish-English bilinguals, and a group of children who transitioned from being Spanish monolingual to Spanish-English bilinguals during the course of the study. Compared to their English monolingual peers, bilingual children from Spanish-speaking homes showed higher IC performance at Head Start entry, as well as steeper IC growth over time. Children who were Spanish monolingual at the beginning of Head Start showed the lowest IC performance at baseline. However, their rate of IC growth exceeded that of children who remained English monolingual and did not differ from that of their peers who entered Head Start being bilingual. These results suggest that acquiring bilingualism and continued bilingual experience are associated with more rapid IC development during the transition from preschool to kindergarten in children from lower SES backgrounds.

Trajectories of alcohol, marijuana, and illicit drug use in a diverse sample of young men who have sex with men

INTRODUCTION

Cross-sectional research has found that young men who have sex with men (YMSM) are more likely to engage in heavy drinking and to have higher rates of marijuana and other illicit drug use compared to their heterosexual peers, but considerably less is known about their patterns of substance use over time.

METHODS

In this study, we combined two longitudinal samples of racially diverse YMSM (N=552) and modeled their substance use trajectories from late-adolescence to young adulthood, including their frequency of alcohol use, frequency of marijuana use, and poly-drug use, using piecewise latent curve growth modeling to model change from ages 17-21 and change from ages 22-24.

RESULTS

We found that all three substance use behaviors increased linearly over the adolescent-to-adult transition. The trajectories for all three substance use behaviors were significantly correlated from ages 17-21. Black YMSM had significantly lower growth from ages 17-21 in alcohol, marijuana, and poly-drug use compared to White YMSM. Hispanic/Latino YMSM had significantly higher growth from ages 22-24 in alcohol use but significantly lower growth in poly-drug use compared to White YMSM. YMSM with higher alcohol frequency slopes and YMSM with higher marijuana use slopes were more likely to have alcohol-related and marijuana-related problems, respectively, at the last wave of the study.

CONCLUSIONS

The results of the present study suggest that the transition from adolescence to adulthood for YMSM is a time of increasing and co-varying substance use and may be a critical period for substance use behaviors to grow into substance use problems.

Contributions of Genes and Environment to Developmental Change in Alcohol Use

The precise nature of how genetic and environmental risk factors influence changes in alcohol use (AU) over time has not yet been investigated. Therefore, the aim of the present study is to examine the nature of longitudinal changes in these risk factors to AU from mid-adolescence through young adulthood. Using a large sample of male twins, we compared five developmental models that each makes different predictions regarding the longitudinal changes in genetic and environmental risks for AU. The best-fitting model indicated that genetic influences were consistent with a gradual growth in the liability to AU, whereas unique environmental risk factors were consistent with an accumulation of risks across time. These results imply that two distinct processes influence adolescent AU between the ages of 15-25. Genetic effects influence baseline levels of AU and rates of change across time, while unique environmental effects are more cumulative.

Religious Development in African American Adolescents: Growth Patterns That Offer Protection

Although religiosity promotes resilient outcomes in African American (AA) adolescents, there is a lack of research that examines the protective role of religious development in the context of stressful life events (SLEs). In 1,595 AA adolescents, a cohort-sequential design was used from ages 12 to 18 to examine subtypes of religiosity, as well as distinct developmental patterns within subtypes. The protective role of religious development was also examined in the context of SLEs. The study findings indicated two subtypes of religiosity-extrinsic and intrinsic religiosity-as well as low and high developmental patterns in the identified subtypes. Furthermore, the protective influence of extrinsic and intrinsic religiosity gradually diminished from age 12 to 18 in the context of SLEs.

Multiple imputation as a flexible tool for missing data handling in clinical research

The last 20 years has seen an uptick in research on missing data problems, and most software applications now implement one or more sophisticated missing data handling routines (e.g., multiple imputation or maximum likelihood estimation). Despite their superior statistical properties (e.g., less stringent assumptions, greater accuracy and power), the adoption of these modern analytic approaches is not uniform in psychology and related disciplines. Thus, the primary goal of this manuscript is to describe and illustrate the application of multiple imputation. Although maximum likelihood estimation is perhaps the easiest method to use in practice, psychological data sets often feature complexities that are currently difficult to handle appropriately in the likelihood framework (e.g., mixtures of categorical and continuous variables), but relatively simple to treat with imputation. The paper describes a number of practical issues that clinical researchers are likely to encounter when applying multiple imputation, including mixtures of categorical and continuous variables, item-level missing data in questionnaires, significance testing, interaction effects, and multilevel missing data. Analysis examples illustrate imputation with software packages that are freely available on the internet.

Longitudinal patterns of pain in patients with diffuse and limited systemic sclerosis: integrating medical, psychological, and social characteristics

PURPOSE

Pain is a common but understudied quality of life concern in systemic sclerosis (SSc). This investigation sought to describe patient-reported pain during the early phase of the disease and to examine potential predictors of this over time.

METHODS

A prospective cohort (N = 316) of patients with early-disease SSc from the Genetics versus ENvironment In Scleroderma Outcome Study (GENISOS) were followed for 3 years. Multilevel modeling was used to describe longitudinal changes in pain and the extent to which pain variance was explained by disease type, emotional health, perceived physical health, health worry, and social support.

RESULTS

Patient-reported pain remained relatively stable, with slight improvement over time. More severe disease type was associated with worse initial pain, but the association was reduced to nonsignificance after accounting for the psychosocial variables. Better emotional health and perceived physical health were associated with lower initial pain. There were marginal interactive effects for perceived physical health and social support such that initial perceptions of poorer physical health, and higher social support, were predictive of greater improvements in pain over time.

CONCLUSIONS

These data suggest that emotional health, perceived physical health, and social support are more relevant to longitudinal SSc pain than disease severity and that perceived physical health and social support may impact pain trajectories. Researchers and rheumatology health professionals should consider these factors in comprehensive pain models and pain management protocols.

Sowing seeds for future generations

In this longitudinal study, we examined the relationship between the trajectory of generative concern measured at ages 23, 26 and 32 and environmental narrative identity at age 32. Canadian participants completed a questionnaire on generative concern at ages 23, 26 and 32 and were then interviewed about their personal experiences with the environment at age 32 ( N = 112). Narratives were coded by independent raters for meaning, vividness and impact, with higher levels indicating a more salient environmental narrative identity. Latent growth models revealed significant individual variability in the trajectories of generative concern from ages 23 to 32. This variability was associated with the salience of environmental narrative identity at age 32 through two different developmental processes: (1) having a higher level of generative concern at age 23 predicted a more salient environmental narrative identity at age 32; and (2) those who developed higher levels of generative concern during the course of emerging adulthood (from ages 23 to 32) also appeared to display a more salient environmental narrative identity at age 32. Implications of these findings are also discussed.

Examining the bidirectional relationship between physical activity, screen time, and symptoms of anxiety and depression over time during adolescence

More physical activity (PA) and less screen time (ST) are positively associated with mental health in adolescents; however, research is limited by short-term designs and the exclusion of ST when examining PA. We examined: (a) changes in PA, ST, symptoms of depression, and symptoms of anxiety over four assessments spanning 11years, and (b) bidirectional relationships between initial PA, ST, and symptoms of depression and anxiety as predictors of change in each other during adolescence. Between 2006 and 2010, participants from Ottawa Canada (Time1; N=1160, Mean age=13.54years) completed questionnaires at four points covering the ages from 10 to 21years. Latent growth modeling was used. PA decreased over time whereas ST and symptoms of depression and anxiety increased over time. Controlling for sex, ethnicity, school location, zBMI, birth year, and parents' education, initially higher anxiety was associated with initially higher ST (covariance=.88, p<.05) and initially lower PA (covariance=-6.84, p=.07) independent of initial symptoms of depression. Higher initial depression was associated with higher initial ST (covariance=2.55, p<.05). Increases in anxiety were associated with increases in ST (covariance=.07, p=.06) and increases in depression (covariance=.41, p<.05). Examining bidirectional relationships, higher initial symptoms of depression predicted greater decreases in PA (b=-.28, p<.05). No other significant findings between initial PA, ST, anxiety, or depression were found as predictors of change in each other. Interventions targeting depression around age 13 may be useful to prevent further declines in PA. Similarly, interventions to reduce ST may be beneficial for concurrent reductions in symptoms of depression and anxiety, irrespective of PA.

Nonlinear latent curve models for multivariate longitudinal data

Latent curve models have become a useful approach to analyzing longitudinal data, due in part to their allowance of and emphasis on individual differences in features that describe change. Common applications of latent curve models in developmental studies rely on polynomial functions, such as linear or quadratic functions. Although useful for describing linear forms of change and some that are nonlinear, latent curve models based on polynomial functions are not suitable for describing many developmental processes that change in a nonlinear manner. This article considers nonlinear latent curve models that permit researchers to consider a variety of nonlinear functions to characterize developmental processes. An example is provided that considers simultaneous development of two behaviors.

The accuracy of adult narrative reports of developmental trajectories

This methodological investigation examines the accuracy of narrative-based scaled ratings covering several post high school years. Guided narratives by young adults described developmentally relevant behaviour and context for each month between ages 17 and the mid-20s. “Prospective” narratives covered shorter time periods in three interviews separated by about 1 year each. A fourth “retrospective” interview included the entire period covered in the previous narratives and took place 1 year after the last prospective interview. Study variables were reliable ratings of data from these carefully conducted and blindly repeated narratives. Aspects of the study design expected to maximise reliability and validity of these data are provided. Prospective–retrospective (test–retest) correlation of ratings based on data from 149 participants covering an average of 64 studied months each on role-related, qualitative, and social variables are reported. The high consistency of many developmental variables based on these reports suggests that retrospective narratives can produce reliable and valid scaled measures covering a substantial period of time.

Does Childhood Attention-Deficit/Hyperactivity Disorder (ADHD) Predict Levels of Depressive Symptoms during Emerging Adulthood?

Little is known about the development and course of depressive symptoms through emerging adulthood among individuals with a childhood history of attention-deficit/hyperactivity disorder (ADHD). The aim of this study was to examine if a history of ADHD in childhood significantly predicted depressive symptoms during emerging adulthood (i.e., ages 18-25 years), including the initial level of depressive symptoms, continued levels of depressive symptoms at each age year, and the rate of change in depressive symptoms over time. 394 participants (205 with ADHD and 189 without ADHD; 348 males and 46 females) drawn from the Pittsburgh ADHD Longitudinal Study (PALS) completed annual self-ratings of depressive symptoms between the ages of 18 and 25 years. Childhood history of ADHD significantly predicted a higher initial level of depressive symptoms at age 18, and higher levels of depressive symptoms at every age year during emerging adulthood. ADHD did not significantly predict the rate of change in depressive symptoms from age 18 to age 25. Childhood history of ADHD remained a significant predictor of initial level of depressive symptoms at age 18 after controlling for comorbid psychiatric diagnoses, but not after controlling for concurrent ADHD symptoms and psychosocial impairment. Participants with childhood histories of ADHD experienced significantly higher levels of depressive symptoms than non-ADHD comparison participants by age 18 and continued to experience higher, although not increasing, levels of depressive symptoms through emerging adulthood. Clinical implications and directions for future research are discussed.

Should Parents Allow Their Adolescent Children to Drink at Home? Family Factors as Predictors of Alcohol Involvement Trajectories Over 15 Years

OBJECTIVE

The present study examined familial risk and protective factors as moderators of parents allowing their adolescent children to drink at home on longitudinal alcohol involvement trajectories.

METHOD

A total of 772 community adolescents and their parents provided data beginning in 1989 and at four subsequent time points over 15 years; Black adolescents were intentionally oversampled (50% at baseline).

RESULTS

Outcomes related to allowing adolescents to drink at home depended on family structure: Adolescents from intact families who were allowed to drink at home showed the lowest levels of alcohol use and problems over time, whereas those from nonintact families who were allowed to drink at home showed the highest levels of involvement. These results controlled for family history of alcohol problems, consistent parenting styles, and demographic characteristics.

CONCLUSIONS

Results suggest that allowing adolescents to drink at home is neither inherently protective nor risky but depends on the family context. Implications for the development of adolescent alcohol involvement are discussed.

Forms of Attrition in a Longitudinal Study of Religion and Health in Older Adults and Implications for Sample Bias

The use of longitudinal designs in the field of religion and health makes it important to understand how attrition bias may affect findings in this area. This study examines attrition in a 4-wave, 8-year study of older adults. Attrition resulted in a sample biased toward more educated and more religiously involved individuals. Conditional linear growth curve models found that trajectories of change for some variables differed among attrition categories. Ineligibles had worsening depression, declining control, and declining attendance. Mortality was associated with worsening religious coping styles. Refusers experienced worsening depression. Nevertheless, there was no evidence of bias in the key religion and health results.

Does Growth in the Executive System of Working Memory Underlie Growth in Literacy for Bilingual Children With and Without Reading Disabilities?

This cohort-sequential study explored the components of working memory (WM) that underlie second language (L2) reading growth in 450 children at risk and not at risk for reading disabilities (RD) whose first language is Spanish. English language learners designated as balanced and nonbalanced bilinguals with and without risk for RD in Grades 1, 2, and 3 at Wave 1 were administered a battery of cognitive (short-term memory, WM, naming speed, and inhibition), vocabulary, and reading measures in Spanish and English. These same measures were administered 1 and 2 years later. Two important findings occurred: First, growth in the WM executive component was significantly related to growth in English word identification and passage comprehension when competing measures (phonological processing, naming speed, inhibition, and fluid intelligence) were entered into the multilevel growth model. Second, children defined as at risk for RD in Wave 1 had lower intercepts than children not at risk at Wave 3 across several measures of cognition, language, and achievement. However, except on measures of the executive component of WM, no significant group differences in linear growth emerged. These findings suggest that growth in L2 reading was tied to growth in the executive system of WM.

Bidirectional associations between alcohol use and sexual risk-taking behavior from adolescence into young adulthood

Overwhelming evidence indicates that sexual risk-taking behavior and alcohol use are linked, but the nature, strength, and timing of these relations may differ between gender and racial subgroups. These issues were addressed by examining the course and interrelations of both behaviors from adolescence into young adulthood, as well as how these patterns differed between both men and women and between Blacks and Whites. Data came from a representative, community-based sample of 1867 urban participants surveyed up to 5 times over a 15-year period. Although both prospective and trajectory analyses showed that adolescent involvement in one behavior predicted later involvement in the other, most patterns were moderated by gender, race, or both. In general, positive, bidirectional associations were discovered among men and Whites. Among women, adolescent sexual risk-taking behavior positively predicted later drinking, but not vice versa. For Blacks, adolescent alcohol use was inconsistently related to later sexual risk-taking behavior, and adolescent sexual risk-taking negatively predicted later alcohol use. Results suggest that associations between sexual risk-taking behavior and alcohol use are more complex than previously thought and that an adequate understanding of these links must account for both gender and racial differences.

Testing Models for the Contributions of Genes and Environment to Developmental Change in Adolescent Depression

We tested two models to identify the genetic and environmental processes underlying longitudinal changes in depression among adolescents. The first assumes that observed changes in covariance structure result from the unfolding of inherent, random individual differences in the overall levels and rates of change in depression over time (random growth curves). The second assumes that observed changes are due to time-specific random effects (innovations) accumulating over time (autoregressive effects). We found little evidence of age-specific genetic effects or persistent genetic innovations. Instead, genetic effects are consistent with a gradual unfolding in the liability to depression and rates of change with increasing age. Likewise, the environment also creates significant individual differences in overall levels of depression and rates of change. However, there are also time-specific environmental experiences that persist with fidelity. The implications of these differing genetic and environmental mechanisms in the etiology of depression are considered.

Ignoring individual differences in times of assessment in growth curve modeling

Researchers often collect longitudinal data to model change over time in a phenomenon of interest. Inevitably, there will be some variation across individuals in specific time intervals between assessments. In this simulation study of growth curve modeling, we investigate how ignoring individual differences in time points when modeling change over time relates to convergence and admissibility of solutions, bias in estimates of parameters, efficiency, power to detect change over time, and Type I error rate. We manipulated magnitude of the individual differences in assessment times, distribution of assessment times, magnitude of change over time, number of time points, and sample size. In contrast to the correct analysis, ignoring individual differences in time points frequently led to inadmissible solutions, especially with few time points and small samples, regardless of the specific magnitude of individual differences that were ignored. Mean intercept and slope were generally estimated without bias. Ignoring individual differences in time points sometimes yielded overestimated intercept and slope variances and underestimated intercept–slope covariance and residual variance. Parameter efficiency as well as power and Type I error rates for the linear slope were unaffected by the type of analysis.

Developmental trajectories of anxiety symptoms in early adolescence: the influence of anxiety sensitivity

Children and adolescents seem to suffer from anxiety disorders at rates similar to adults. Interestingly, anxiety symptoms appear to generally decline over time within children as evidenced by lower rates in early and middle adolescence. There is some evidence that there may be heterogeneous subpopulations of adolescent children with different trajectories of anxiety symptoms, including a class of adolescents with elevated levels of anxiety that do not dissipate over time. Anxiety sensitivity has been identified as an important risk factor in the development of anxiety psychopathology. This study prospectively examined the development of anxiety symptoms in a sample of 277 adolescents (M age = 11.52; 44 % female, 56 % male) over a 3 year period including the influence of anxiety sensitivity on this development. Further, this study investigated whether there were distinct classes of adolescents based on their anxiety symptom trajectories and including anxiety sensitivity as a predictor. Consistent with other reports, findings indicated an overall decline in anxiety symptoms over time in the sample. However, three classes of adolescents were found with distinct anxiety symptom trajectories and anxiety sensitivity was an important predictor of class membership. Adolescents with elevated anxiety sensitivity scores were more likely to be classified as having high and increasing anxiety symptoms over time versus having moderate to low and decreasing anxiety symptoms over time. There are important implications for identification of adolescents and children who are at risk for the development of an anxiety disorder.

Genetic influences on alcohol use across stages of development: GABRA2 and longitudinal trajectories of drunkenness from adolescence to young adulthood

Longitudinal analyses allow us to understand how genetic risk unfolds across development, in a way that is not possible with cross-sectional analyses of individuals at different ages. This has received little attention in genetic association analyses. In this study, we test for genetic effects of GABRA2, a gene previously associated with alcohol dependence, on trajectories of drunkenness from age 14 to 25. We use data from 1070 individuals who participated in the prospective sample of the Collaborative Study on the Genetics of Alcoholism, in order to better understand the unfolding of genetic risk across development. Piecewise linear growth models were fit to model the influence of genotype on rate of increase in drunkenness from early adolescence to young adulthood (14-18 years), the change in drunkenness during the transition to adulthood (18-19 years) and the rate of change in drunkenness across young adulthood (≥ 19 years). Variation in GABRA2 was associated with an increase in drunkenness that occurred at the transition between adolescence and adulthood. The genotypic effect was more pronounced in females. These analyses illustrate the importance of longitudinal data to characterize how genetic effects unfold across development. The findings suggest that transitions across important developmental periods may alter the relative importance of genetic effects on patterns of alcohol use. The findings also suggest the importance of considering gender when evaluating genetic effects on drinking patterns in males and females.

An idiographic approach to estimating models of dyadic interactions with differential equations

We present an idiographic approach to modeling dyadic interactions using differential equations. Using data representing daily affect ratings from romantic relationships, we examined several models conceptualizing different types of dyadic interactions. We fitted each model to each of the dyads and the resulting AICc values were used to classify the most likely configuration of interaction for each dyad. Additionally, the AICc from the different models were used in parameter averaging across models. Averaged parameters were used in models involving predictors of relationship dynamics, as indexed by these parameters, as well as models wherein the parameters predicted distal outcomes of the dyads such as relationship satisfaction and status. Results indicated that, within our sample, the most likely interaction style was that of independence, without evidence of emotional interrelations between the two individuals in the couple. Attachment-related avoidance and anxiety showed significant relations with model parameters, such that ideal levels of affect for males were negatively influenced by higher levels of avoidance from their partner while their own levels of anxiety had positive effects on their levels of dyadic coregulation. For females coregulation was negatively influenced by both time in the relationship and their partner's level of avoidance. Analysis involving distal outcomes showed modest influences from the individual's level of ideal affect.

Exposure to stressors and trajectories of perceived stress among older adults

OBJECTIVES

Models of stress incorporate both the environmental demands experienced by individuals (stressors) and the appraisal of these life events (perceptions). Because little is known about the extent to which experience and perceptions are related, we examine this relationship in a nationally representative population of older Taiwanese adults.

METHOD

Using growth models applied to data from 3 waves (1999, 2003, and 2007) of the Taiwan Longitudinal Study of Aging, we (a) investigate patterns of change in perceived stress in later adulthood and (b) examine how experienced stressors influence perceived stress. Participants were asked to report the presence of, and in some cases the degree of, exposure to stressors including total number of medical conditions, difficulty with activities of daily living, difficulty with mobility functions, being financially worse off compared with the prior wave, experiencing the death of a child, and experiencing a marital disruption. Items reflecting perceived stress included concerns about various domains pertaining to the respondent and his/her family member.

RESULTS

Our results indicate that exposure to stressors increases, whereas perceived stress decreases, over time. Change in exposure to stressors is not generally associated with change in perceptions of stress, with the exception of a summary measure of health-related exposure to stressors. An increase in poor health over time is related to an increase in perceived stress in all domains.

DISCUSSION

The results underscore the importance of distinguishing between perceptions of stress and exposure to stressors when studying the links between stress and health among older adults. Furthermore, the diminishing linkage between experienced stressors and perceptions of stress suggests that older adults' appraisal may be an adaptive coping strategy that emerges to buffer some of the difficulties that are inevitable in later life.

A Moderated Nonlinear Factor Model for the Development of Commensurate Measures in Integrative Data Analysis

Integrative data analysis (IDA) is a methodological framework that allows for the fitting of models to data that have been pooled across two or more independent sources. IDA offers many potential advantages including increased statistical power, greater subject heterogeneity, higher observed frequencies of low base-rate behaviors, and longer developmental periods of study. However, a core challenge is the estimation of valid and reliable psychometric scores that are based on potentially different items with different response options drawn from different studies. In Bauer and Hussong (2009) we proposed a method for obtaining scores within an IDA called moderated nonlinear factor analysis (MNLFA). Here we move significantly beyond this work in the development of a general framework for estimating MNLFA models and obtaining scale scores across a variety of settings. We propose a five step procedure and demonstrate this approach using data drawn from n=1972 individuals ranging in age from 11 to 34 years pooled across three independent studies to examine the factor structure of 17 binary items assessing depressive symptomatology. We offer substantive conclusions about the factor structure of depression, use this structure to compute individual-specific scale scores, and make recommendations for the use of these methods in practice.

Growth trajectories of alcohol information processing and associations with escalation of drinking in early adolescence

This longitudinal study provided a comprehensive examination of age-related changes in alcohol outcome expectancies, subjective evaluation of alcohol outcomes, and automatic alcohol associations in early adolescence. A community sample (52% female, 75% White/non-Hispanic) was assessed annually for 3 years (mean age at the first assessment = 11.6 years). Results from growth modeling suggested that perceived likelihood of positive outcomes increased and that subjective evaluations of these outcomes were more positive with age. Perceived likelihood of negative outcomes declined with age. Automatic alcohol associations were assessed with an Implicit Association Task (IAT), and were predominantly negative, but these negative associations weakened with age. High initial levels of perceived likelihood of positive outcomes at age 11 were associated with escalation of drinking. Perceived likelihood of negative outcomes was associated with low risk for drinking at age 11, but not with changes in drinking. Increases in positive evaluations of positive outcomes were associated with increases in alcohol use. Overall, findings suggest that at age 11, youth maintain largely negative attitudes and perceptions about alcohol, but with the transition into adolescence, there is a shift toward a more neutral or ambivalent view of alcohol. Some features of this shift are associated with escalation of drinking. Our findings point to the importance of delineating multiple aspects of alcohol information processing for extending cognitive models of alcohol use to the early stages of drinking.

Prenatal cocaine exposure and trajectories of externalizing behavior problems in early childhood: examining the role of maternal negative affect

This study examined the association between prenatal cocaine exposure (PCE) and developmental trajectories of externalizing behavior problems from 18 to 54 months of child age. A hypothesized indirect association between PCE and externalizing trajectories via maternal negative affect was also examined. Caregiving environmental risk and child sex were evaluated as moderators. This study consisted of 196 mother-child dyads recruited at delivery from local area hospitals (107 PCE, 89 non-PCE) and assessed at seven time points across the toddler to preschool periods. Results revealed no direct associations between PCE and externalizing behavior problem trajectories. However, results did indicate that PCE shared a significant indirect relationship with externalizing behavior problem trajectories via higher levels of maternal negative affect. The association between PCE and externalizing problem trajectories was also moderated by caregiving environmental risk such that PCE children in high-risk caregiving environments did not experience the well-documented normative decline in externalizing behavior problems beginning at around 3 years of age. This study suggests potential pathways to externalizing behavior problems among high-risk children.

Latent growth modeling with domain-specific outcomes comprised of mixed response types in intervention studies

OBJECTIVE

When several continuous outcome measures of interest are collected across time in experimental studies, the use of standard statistical procedures, such as multivariate analysis of variance or growth curve modeling, can be properly used to assess treatment effects. However, when data consist of mixed responses (e.g., continuous and ordered categorical [ordinal] responses), traditional modeling approaches are no longer appropriate. The purpose of this article is to illustrate the use of a more suitable modeling procedure when mixed responses are collected in longitudinal intervention studies.

METHOD

Problems with traditional analyses of such data are discussed, as are potential advantages provided by the proposed modeling approach. The application of the multiple-domain latent growth modeling approach with mixed responses is illustrated for experimental designs with data from the SeniorWISE study (McDougall et al., 2010). This multisite randomized trial assessed memory functioning of 265 elderly adults across a 26-month period after receiving either a memory or health promotion training program.

RESULTS

The latent growth models illustrated allow one to examine treatment effects on the growth of multiple mixed outcomes while incorporating associations among multiple responses, which allows for better missing data treatment, greater power, and more accurate control of Type I error. The interpretation of parameters of interest and treatment effects is discussed using the SeniorWISE data.

CONCLUSIONS

Multiple-domain latent growth modeling with mixed responses is a flexible statistical modeling tool that can have substantial benefits for applied researchers. As such, the use of this modeling approach is expected to increase.

The dynamic role of genetics on cortical patterning during childhood and adolescence

Longitudinal imaging and quantitative genetic studies have both provided important insights into the nature of human brain development. In the present study we combine these modalities to obtain dynamic anatomical maps of the genetic contributions to cortical thickness through childhood and adolescence. A total of 1,748 anatomic MRI scans from 792 healthy twins and siblings were studied with up to eight time points per subject. Using genetically informative latent growth curve modeling of 81,924 measures of cortical thickness, changes in the genetic contributions to cortical development could be visualized across the age range at high resolution. There was highly statistically significant (P < 0.0001) genetic variance throughout the majority of the cerebral cortex, with the regions of highest heritability including the most evolutionarily novel regions of the brain. Dynamic modeling of changes in heritability over time demonstrated that the heritability of cortical thickness increases gradually throughout late childhood and adolescence, with sequential emergence of three large regions of high heritability in the temporal poles, the inferior parietal lobes, and the superior and dorsolateral frontal cortices.

Optimal assignment methods in three-form planned missing data designs for longitudinal panel studies

Planned missing designs are becoming increasingly popular, but because there is no consensus on how to implement them in longitudinal research, we simulated longitudinal data to distinguish between strategies of assigning items to forms and of assigning forms to participants across measurement occasions. Using relative efficiency as the criterion, results indicate that balanced item assignment coupled with assigning different forms over time most often yields the optimal assignment method, but only if variables are reliable. We also address how practice effects can bias latent means. A second simulation demonstrates that (a) assigning different forms over time diminishes practice effects and (b) using planned-missing-data patterns as predictors of practice can remove bias altogether.

The Consequences of Ignoring Variability in Measurement Occasions Within Data Collection Waves in Latent Growth Models

In longitudinal data collection, it is common that each wave of collection spans several months. However, researchers using latent growth models commonly ignore variability in data collection occasions within a wave. In this study, we investigated the consequences of ignoring within-wave variability in measurement occasions using a Monte Carlo simulation and an empirical study. The results of the simulation study showed that ignoring heterogeneity resulted in biased estimates for some parameters, especially when heterogeneity was large and assessment dates had a skewed distribution. Models constructed on person-specific time points yielded precise estimates and more adequate model fit. In the empirical study, we demonstrated different time coding strategies with a subsample taken from Early Childhood Longitudinal Study Kindergarten Cohort.

Rebound sex: Sexual motives and behaviors following a relationship breakup

The present study used a longitudinal, online diary method to examine trajectories of psychological recovery and sexual experience following a romantic relationship breakup among 170 undergraduate students. Consistent with popular beliefs about rebound and revenge sex, having sex to cope with distress and to get over or get back at the ex-partner were elevated immediately following the breakup and then declined over time, as did the probability of having sex with a new partner. Also consistent with popular lore, those who were "dumped" by their partners were more distressed and angry and more likely to have sex to cope and to get back at or get over their ex-partner. Finally, individuals who reported having sex to cope with negative feelings or to get over their ex-partner at the beginning of the study were more likely to have sex with a stranger and to continue having sex with new partners over time. Results were discussed in terms of widely held but largely untested beliefs about rebound and revenge sex.

Dyadic Curve-of-Factors Model: An Introduction and Illustration of a Model for Longitudinal Non-Exchangeable Dyadic Data

The analysis of longitudinal data collected from non-exchangeable dyads presents a challenge for applied researchers for various reasons. This paper introduces the Dyadic Curve-of-Factors Model (D-COFM) which extends the Curve-of-Factors Model (COFM) proposed by McArdle (1988) for use with non-exchangeable dyadic data. The D-COFM overcomes problems with modeling composite scores across time and instead permits examination of the growth in latent constructs over time. The D-COFM also appropriately models the interdependency among non-exchangeable dyads. Different parameterizations of the D-COFM are illustrated and discussed using a real dataset to aid applied researchers when analyzing dyadic longitudinal data.

Longitudinal mediation of processing speed on age-related change in memory and fluid intelligence

Age-related decline in processing speed has long been considered a key driver of cognitive aging. While the majority of empirical evidence for the processing speed hypothesis has been obtained from analyses of between-person age differences, longitudinal studies provide a direct test of within-person change. Using recent developments in longitudinal mediation analysis, we examine the speed-mediation hypothesis at both the within-and between-person levels in two longitudinal studies, Longitudinal Aging Study Amsterdam (LASA) and Origins of Variance in the Oldest-Old (OCTO-Twin). We found significant within-person indirect effects of change in age, such that increasing age was related to lower speed, which in turn relates to lower performance across repeated measures on other cognitive outcomes. Although between-person indirect effects were also significant in LASA, they were not in OCTO-Twin which is not unexpected given the age homogeneous nature of the OCTO-Twin data. A more in-depth examination through measures of effect size suggests that, for the LASA study, the within-person indirect effects were small and between-person indirect effects were consistently larger. These differing magnitudes of direct and indirect effects across levels demonstrate the importance of separating between- and within-person effects in evaluating theoretical models of age-related change.

Trajectories of depressive symptoms and suicidality among heterosexual and sexual minority youth

Sexual minority youth report higher rates of depression and suicidality than do heterosexual youth. Little is known, however, about whether these disparities continue as youth transition into young adulthood. The primary goals of this study were to describe and compare trajectories of adolescent depressive symptoms and suicidality among sexual minority and heterosexual youth, examine differences in depressive symptoms and suicidality trajectories across sexual orientation subgroups, and determine whether there are gender differences in these longitudinal disparities. Four waves of data from the National Longitudinal Study of Adolescent Health were analyzed using latent curve modeling (N = 12,379; 53% female). Results showed that the rates of depressive symptoms and suicidality in early adolescence were higher among sexual minority youth than among heterosexual youth, and that these disparities persisted over time as participants transitioned into young adulthood. Consistent with previous cross-sectional studies, the observed longitudinal disparities were largest for females and for bisexually-identified youth. Sexual minority youth may benefit from childhood and early adolescent prevention and intervention programs.

The moderating effect of family involvement on substance use risk factors in adolescents with severe emotional and behavioral challenges

OBJECTIVE

The current study longitudinally examined the moderating effects of family involvement, which previous research has shown to be a protective factor against adolescents' substance use involvement, on the associations between internalizing and externalizing problems, respectively, and substance use (alcohol, marijuana, and tobacco) and problems in a unique sample of adolescents with severe emotional and behavioral problems.

METHOD

Adolescents (n=4786) and their parents were assessed at three waves over a 1-year period as part of the SAMHSA funded Comprehensive Community Mental Health Initiative (CMHI). Multilevel growth curve models were estimated using Hierarchical Linear Modeling (HLM) to examine the associations between time-varying internalizing and externalizing problems, respectively, and substance use involvement over time as a function of family involvement at baseline.

RESULTS

Results supported the hypothesis that family involvement protects against adolescent substance use involvement by buffering the adverse effects of both internalizing and externalizing problems. Specifically, for alcohol use, family involvement buffered the negative effects of high externalizing problems. For tobacco use, family involvement buffered the negative effects of both internalizing and externalizing problems, respectively. For substance use problems, family involvement buffered the effects of only internalizing problems. Family involvement did not moderate effects of emotional and behavioral problems on marijuana use, however, a significant main effect was observed such that family involvement was negatively related to marijuana use.

CONCLUSIONS

Results suggest that the beneficial effects of family involvement on at-risk adolescents' substance use involvement are dependent on the type of emotional and behavioral problems the adolescent experiences as well as the type of substance the adolescent uses. Implications for family- and adolescent-focused treatment are discussed.

Investigating terminal decline: results from a UK population-based study of aging

The terminal decline hypothesis states that in the proximity of death, an individual's decline in cognitive abilities accelerates. We aimed at estimating the onset of faster rate of decline in global cognition using Mini Mental State Examination (MMSE) scores from participants of the Cambridge City over 75 Cohort Study (CC75C), a U.K. population-based longitudinal study of aging where almost all participants have died. The random change point model fitted to MMSE scores structured as a function of distance to death allowed us to identify a potentially different onset of change in rate of decline before death for each individual in the sample. Differences in rate of change before and after the onset of change in rate of decline by sociodemographic variables were investigated. On average, the onset of a faster rate of change occurred about 7.7 years before death and varied across individuals. Our results show that most individuals experience a period of slight decline followed by a much sharper decline. Education, age at death, and cognitive impairment at study entry were identified as modifiers of rate of change before and after change in rate of decline. Gender differences were found in rate of decline in the final stages of life. Our study suggests that terminal decline is a heterogeneous process, with its onset varying between individuals.

Codevelopment of externalizing and internalizing symptoms in middle to late childhood: sex, baseline respiratory sinus arrhythmia, and respiratory sinus arrhythmia reactivity as predictors

We investigated the roles of sex and respiratory sinus arrhythmia (RSA), an index of autonomic parasympathetic nervous system activity, as predictors of codeveloping externalizing and internalizing symptoms in middle childhood. We expected that sex, baseline RSA (RSA-B), and RSA reactivity (RSA-R) to two types of tasks would interact to differentiate co-occurring trajectories of symptoms. We tested these hypotheses by combining longitudinal data from two independent samples (n = 390; 210 girls, 180 boys) with repeated measures at ages 8, 9, 10, and 11. RSA-R was measured in response to a socially stressful and frustrating stressor. Indicators of growth in externalizing and internalizing symptoms were derived from multiple domain growth models and used in person-centered growth mixture analyses. Three groups of externalizing and internalizing trajectories were found. Profile membership was predicted by several two-way interactions among sex, RSA-B, or RSA-R but was not predicted by three-way interactions. Children with low RSA-B and strong RSA withdrawal, girls with low RSA-B, and girls with strong RSA withdrawal were more likely to be on a developmental trajectory of low externalizing symptoms and moderately elevated internalizing symptoms. Membership in the high externalizing and high internalizing trajectory was predicted by weak RSA withdrawal for boys and strong RSA withdrawal for girls. The type of stressor task also played a role in predicting probability of profile membership. Results are discussed in the context of developmental psychobiology and implications for the codevelopment of psychopathology symptoms in childhood.

Voluntary leadership roles in religious groups and rates of change in functional status during older adulthood

Linear growth curve modeling was used to compare rates of change in functional status between three groups of older adults: Individuals holding voluntary lay leadership positions in a church, regular church attenders who were not leaders, and those not regularly attending church. Functional status was tracked longitudinally over a 4-year period in a national sample of 1,152 Black and White older adults whose religious backgrounds were either Christian or unaffiliated. Leaders had significantly slower trajectories of increase in both the number of physical impairments and the severity of those impairments. Although regular church attenders who were not leaders had lower mean levels of impairment on both measures, compared with those not regularly attending church, the two groups of non-leaders did not differ from one another in their rates of impairment increase. Leadership roles may contribute to longer maintenance of physical ability in late life, and opportunities for voluntary leadership may help account for some of the health benefits of religious participation.

Evaluating a bilingual voluntary community-based healthcare organization

The current study compared traditional recovery homes for individuals with substance use disorders with homes that had been modified to feature culturally congruent communication styles. Findings indicated significant increases in employment income, with the size of the change significantly greater in the culturally modified houses. Significant decreases in alcohol use over time were also found, with larger decreases over time in the traditional recovery homes. Use of prescribed medications and days using drugs significantly decreased over time, but not differentially for those in the two types of recovery homes. The implications of these findings are discussed.

Distinguishing between-person and within-person relationships in longitudinal health research: arthritis and quality of life

BACKGROUND

Many health measures (e.g., blood pressure, quality of life) have meaningful fluctuation over time around a relatively stable mean level for each person.

PURPOSE

This didactic paper describes two closely related statistical models for examining between-person and within-person relationships between two or more sets of measures collected over time: the latent intercept model with correlated residuals (LI) in structural equation modeling framework and the multivariate multilevel model (MVML) in multilevel modeling framework.

RESULTS

We illustrated that the basic LI model and the MVML model are equivalent. We presented an illustrative example using a national arthritis data resource to examine between-person and within-person relationships of symptom status, functional health, and quality of life in arthritis patients.

DISCUSSION

Additional design and modeling issues for the treatment of missing data are considered. We discuss contexts in which one of the two models may be preferred. Mplus and SAS syntax are available.

Estimating the Reliability of Aggregated and Within-Person Centered Scores in Ecological Momentary Assessment

A procedure for estimating the reliability of test scores in the context of ecological momentary assessment (EMA) was proposed to take into account the characteristics of EMA measures. Two commonly used test scores in EMA were considered: the aggregated score (AGGS) and the within-person centered score (WPCS). Conceptually, AGGS and WPCS represent the interindividual differences and the intraindividual differences, respectively. The reliability coefficients for AGGS and WPCS were derived using a multilevel factor model with a serial correlation structure framework. Point estimates and confidence intervals of these coefficients were obtained using Mx ( Neale, Boker, Xie, & Maes, 2004 ). A simulation study showed that the proposed procedure performed well empirically. Diary data from Huang (2009) , which recorded daily joy level of 110 undergraduate students for 8 days, was used to illustrate the applicability of the proposed method.