Veronica T. Cole

Delineating prototypical patterns of substance use initiations over time

AIMS The purpose of this paper is to discover patterns of drug use initiations over time through a multiple event process survival mixture model (MEPSUM model), a novel approach for substance use and prevention research. DESIGN The MEPSUM model combines survival analysis and mixture modeling-specifically latent class analysis-to examine individual differences in the timing of initiation and cumulative risk of substance use over time, and is applied to cross-sectional survey data on drug initiations. SETTING Data are drawn from the 2009 National Survey on Drug Use and Health. PARTICIPANTS The survey includes responses from 55 772 individuals (52.05% female). MEASUREMENTS The age of first use of nine different types of substances are examined, including alcohol, tobacco, cocaine and non-medical use of prescription drugs. FINDINGS It is argued that six patterns parsimoniously describe the populations risk of initiating different substances over time, described colloquially as general abstainers; early, late and progressive soft drug users; and early and late hard drug users. Both gender and ethnicity significantly predict the patterns, with Caucasians and males having a higher risk for the hard drug-using patterns. The MEPSUM model produced stable results in this application, as the patterns are validated in a split-sample design. CONCLUSIONS The MEPSUM model provides a statistical framework from which to evaluate patterns of risk for drug initiations over time and predict substance use trajectories relevant to public health interventions. The patterns that result from the model can be used as outcomes for subsequent investigations of etiological and mediating mechanisms.

A Note on the Use of Mixture Models for Individual Prediction

Mixture models capture heterogeneity in data by decomposing the population into latent subgroups, each of which is governed by its own subgroup-specific set of parameters. Despite the flexibility and widespread use of these models, most applications have focused solely on making inferences for whole or subpopulations, rather than individual cases. This article presents a general framework for computing marginal and conditional predicted values for individuals using mixture model results. These predicted values can be used to characterize covariate effects, examine the fit of the model for specific individuals, or forecast future observations from previous ones. Two empirical examples are provided to demonstrate the usefulness of individual predicted values in applications of mixture models. The first example examines the relative timing of initiation of substance use using a multiple event process survival mixture model, whereas the second example evaluates changes in depressive symptoms over adolescence using a growth mixture model.

Improving Factor Score Estimation Through the Use of Observed Background Characteristics

A challenge facing nearly all studies in the psychological sciences is how to best combine multiple items into a valid and reliable score to be used in subsequent modeling. The most ubiquitous method is to compute a mean of items, but more contemporary approaches use various forms of latent score estimation. Regardless of approach, outside of large-scale testing applications, scoring models rarely include background characteristics to improve score quality. This article used a Monte Carlo simulation design to study score quality for different psychometric models that did and did not include covariates across levels of sample size, number of items, and degree of measurement invariance. The inclusion of covariates improved score quality for nearly all design factors, and in no case did the covariates degrade score quality relative to not considering the influences at all. Results suggest that the inclusion of observed covariates can improve factor score estimation.

Parental and adolescent health behaviors and pathways to adulthood

This paper examines associations among parental and adolescent health behaviors and pathways to adulthood. Using data from the National Longitudinal Study of Adolescent to Adult Health, we identify a set of latent classes describing pathways into adulthood and examine health-related predictors of these pathways. The identified pathways are consistent with prior research using other sources of data. Results also show that both adolescent and parental health behaviors differentiate pathways. Parental and adolescent smoking are associated with lowered probability of the higher education pathway and higher likelihood of the work and the work & family pathways (entry into the workforce soon after high school completion). Adolescent drinking is positively associated with the work pathway and the higher education pathway, but decreases the likelihood of the work & family pathway. Neither parental nor adolescent obesity are associated with any of the pathways to adulthood. When combined, parental/adolescent smoking and adolescent drinking are associated with displacement from the basic institutions of school, work, and family.

Advancing the Study of Adolescent Substance Use Through the Use of Integrative Data Analysis

A wealth of information is currently known about the epidemiology, etiology, and evaluation of drug and alcohol use across the life span. Despite this corpus of knowledge, much has yet to be learned. Many factors conspire to slow the pace of future advances in the field of substance use including the need for long-term longitudinal studies of often hard-to-reach subjects who are reporting rare and episodic behaviors. One promising option that might help move the field forward is integrative data analysis (IDA). IDA is a principled set of methodologies and statistical techniques that allow for the fitting of statistical models to data that have been pooled across multiple, independent samples. IDA offers a myriad of potential advantages including increased power, greater coverage of rare behaviors, more rigorous psychometric assessment of theoretical constructs, accelerated developmental time period under study, and enhanced reproducibility. However, IDA is not without limitations and may not be useful in a given application for a variety of reasons. The goal of this article is to describe the advantages and limitations of IDA in the study of individual development over time, particularly as it relates to trajectories of substance use. An empirical example of the measurement of polysubstance use is presented and this article concludes with recommendations for practice.

Recovering Predictor–Criterion Relations Using Covariate-Informed Factor Score Estimates

Although it is currently best practice to directly model latent factors whenever feasible, there remain many situations in which this approach is not tractable. Recent advances in covariate-informed factor score estimation can be used to provide manifest scores that are used in second-stage analysis, but these are currently understudied. Here we extend our prior work on factor score recovery to examine the use of factor score estimates as predictors both in the presence and absence of the same covariates that were used in score estimation. Results show that whereas the relation between the factor score estimates and the criterion are typically well recovered, substantial bias and increased variability is evident in the covariate effects themselves. Importantly, using covariate-informed factor score estimates substantially, and often wholly, mitigates these biases. We conclude with implications for future research and recommendations for the use of factor score estimates in practice.

Depressive Symptomology as a Moderator of Friend Selection and Influence on Substance Use Involvement: Estimates from Grades 6 to 12 in Six Longitudinal School-Based Social Networks

Although the contributions of friend selection and friend influence to adolescent homophily on substance use behaviors has been of enduring research interest, moderators of these processes have received relatively little research attention. Identification of factors that dampen or amplify selection and influence on substance use behaviors is important for informing prevention efforts. Whereas prior research has examined adolescent drinking, smoking, and marijuana use, the current study examined whether friend selection and friend influence operated on substance use involvement, an indicator of problematic use, and whether depressive symptomology moderated these processes. In addition, it examined whether these relationships varied from grade 6 to 12. The study used a cohort-sequential design in which three cohorts of youth (first surveyed in grades, 6, 7, and 8) in six school-based longitudinal social networks were surveyed up to seven times, yielding N = 6817 adolescents (49% female). Stochastic actor-oriented models were applied to test hypothesized relationships in the six networks, then results were synthesized in a meta-analysis. Depressive symptoms did not moderate selection or influence on substance use involvement at any grade level, but indirectly contributed to diffusion of substance use involvement through school networks via patterns of network ties. Research is needed on contextual factors, particularly in schools, that might account for when, if at all, depressive symptoms condition friend selection and influence on substance use.

Simplifying the implementation of modern scale scoring methods with an automated R package: Automated moderated nonlinear factor analysis (aMNLFA)

When generating scores to represent latent constructs, analysts have a choice between applying psychometric approaches that are principled but that can be complicated and time-intensive versus applying simple and fast, but less precise approaches, such as sum or mean scoring. We explain the reasons for preferring modern psychometric approaches: namely, use of unequal item weights and severity parameters, the ability to account for local dependence and differential item functioning, and the use of covariate information to more efficiently estimate factor scores. We describe moderated nonlinear factor analysis (MNLFA), a relatively new, highly flexible approach that allows analysts to develop precise factor score estimates that address limitations of sum score, mean score, and traditional factor analytic approaches to scoring. We then outline the steps involved in using the MNLFA scoring approach and discuss the circumstances in which this approach is preferred. To overcome the difficulty of implementing MNLFA models in practice, we developed an R package, aMNLFA, that automates much of the rule-based scoring process. We illustrate the use of aMNLFA with an empirical example of scoring alcohol involvement in a longitudinal study of 6998 adolescents and compare performance of MNLFA scores with traditional factor analysis and sum scores based on the same set of 12 items. MNLFA scores retain more meaningful variation than other approaches. We conclude with practical guidelines for scoring.

An Empirical Assessment of the Sensitivity of Mixture Models to Changes in Measurement

This study explored the extent to which variations in self-report measures across studies can produce differences in the results obtained from mixture models. Data (N = 854) come from a laboratory analogue study of methods for creating commensurate scores of alcohol- and substance-use-related constructs when items differ systematically across participants for any given measure. Items were manipulated according to 4 conditions, corresponding to increasing levels of alteration to item stems, response options, or both. In Study 1, results from latent class analyses (LCAs) of alcohol consequences were compared across the 4 conditions, revealing differences in class enumeration and configuration. In Study 2, results from factor mixture models (FMMs) of alcohol expectancies were compared across 2 of the conditions, revealing differences in patterns and magnitude of the factor loadings and thresholds. The results suggest that even subtle differences in measurement can have substantively meaningful effects on mixture model results.

Adapting Latent Profile Analysis to Take Into Account Measurement Noninvariance

Measurement invariance has been explored extensively in models for continuous latent variables, but has been studied less systematically in models for categorical latent variables, including latent class analysis (LCA) and latent profile analysis (LPA), among others. In the current analysis, a Monte Carlo simulation was conducted to assess the robustness of LPA to measurement noninvariance. Data in all cells were simulated according to a two-class LPA (N = 500), with varying levels of (a) class separation (two levels: entropy .7 and entropy .9), (b) class membership proportions (two levels: .5/.5 and .8/.2), (c) type of measurement noninvariance (three levels: noninvariant loadings, noninvariant intercepts, and noninvariant loadings and intercepts), and (d) magnitude of noninvariance effects (two levels: large and small). Two sorts of models were fit to the data: (a) a misspecified LPA that includes only covariate effects on latent class membership and (b) a correctly specified model in which partial noninvariance is modeled through direct effects of covariates on indicators. Bias in model parameters, including class-specific means and regression coefficients relating class membership to covariates, as well as modal estimates of class membership, were assessed. Failing to consider measurement noninvariance was linked to bias in parameter estimates as well as modal class membership estimates, particularly given low class separation and large DIF. Modal class membership estimates were strongly affected by intercept noninvariance in these conditions; whereas average misclassification rates were below 10% in the full model, over 20% of cases were often misclassified when only covariate effects on class were included. The same simulated data were then used to test the efficacy of standard itemwise testing algorithms for detecting partial noninvariance. Itemwise tests had generally high sensi-