Covariance Models For Latent Structure In Longitudinal Data

We present several approaches to modeling latent structure in longitudinal studies when the covariance itself is the primary focus of the analysis. This is a departure from much of the work on longitudinal data analysis, in which attention is focused solely on the cross-sectional mean and the influence of covariates on the mean. Such analyses are particularly important in policy-related studies, in which the heterogeneity of the population is of interest. We describe several traditional approaches to this modeling and introduce a flexible, parsimonious class of covariance models appropriate to such analyses. This class, while rooted in the tradition of mixed effects and random coefficient models, merges several disparate modeling philosophies into what we view as a hybrid approach to longitudinal data modeling. We discuss the implications of this approach and its alternatives especially on model interpretation. We compare several implementations of this class to more commonly employed mixed effects models to describe the strengths and limitations of each. These alternatives are compared in an application to long-term trends in wage inequality for young workers. The findings provide additional guidance for the model formulation process in both statistical and substantive senses.     

Sexual Satisfaction and Marital Satisfaction During the Early Years of Chinese Marriage: A Three-Wave,Cross-Lagged,Actor–Partner Interdependence Model

Sexuality is an integral component of many intimate relationships, and research has consistently demonstrated a positive association between sexual and marital satisfaction. However, the temporal dynamics of this association remain controversial and understudied with rigorous longitudinal dyadic approaches, and empirical efforts examining this association in non-Western samples remain sparse. Based on three annual waves of data from 268 Chinese couples during the early years of marriage, this study tested a cross-lagged, actor–partner interdependence model examining the association between sexual and marital satisfaction. Results indicated that (a) across three waves, husbands’ earlier sexual satisfaction predicted their later marital satisfaction, rather than the reverse; (b) from Wave 1 to Wave 2, wives’ earlier marital satisfaction predicted their later sexual satisfaction, rather than the reverse, but no association between wives’ sexual and marital satisfaction was found from Wave 2 to Wave 3; and (c) four longitudinal indirect associations linking sexual and marital satisfaction were identified, including three actor associations and one partner association. Such findings shed light on the complexity inherent within the dynamic association between sexual and marital satisfaction over the early years of marriage in the Chinese cultural context.     

She left, he left: how employment and satisfaction affect women's and men's decisions to leave marriages

Studies examining determinants of divorce have largely ignored differences between factors that elevate wives' and husbands' initiation of divorce. The authors use longitudinal data and a latent class model embedded in a competing-risks event history model to assess distinct predictors of wives and husbands leaving marriages. They find that when men are not employed, either spouse is more likely to leave. When wives report better-than-average marital satisfaction, their employment affects neither spouse's exit. However, when wives report below-average marital satisfaction, their employment makes it more likely they will leave. The authors' findings suggest that theories of divorce require "gendering" to reflect asymmetric gender change.     

Regression Analysis of Multivariate Binary Response Variables Using Rasch-Type Models and Finite-Mixture Methods

A model is considered for the regression analysis of multivariate binary data such as repeated-measures data (for example, panel data) or multiple-indicators with measures of some underlying characteristic such as attitude or ability (for example, surveys or tests). The model is related to the usual Rasch model, the usual latent-class model, and other familiar models such as logistic regression. In addition to a regression specification, the model includes parameters that describe heterogeneity not accounted for by the predictors. In contrast to most other approaches, a nonparametric specification of the latent mixing distribution is used, leading to a formulation based on scaled latent classes. We examine the relationship between this model and several other models, give a tractable formulation of the likelihood function and likelihood equations, present an algorithm for maximum-likelihood estimation, and analyze marginal and conditional latent structures. The approach is illustrated with longitudinal data from the German Socioeconomic Panel.     

The longitudinal effects of parental monitoring and self-control on depression in Korean adolescents: A multivariate latent growth approach

The purpose of this study was to examine the longitudinal effects of parental monitoring and self-control on depression. To address this purpose, this study investigated the interrelationships among depression, perceived parental monitoring, and self-control—as well as their developmental changes from the eighth to the twelfth grades—by repeated assessment of 3449 Korean adolescents. The data from Korea Youth Panel Survey (KYPS) of the eighth graders who were followed for five years were analyzed using latent growth modeling. The univariate latent growth models showed that adolescents experience an increase in both parental monitoring and self-control but a decline in depression over the five years. In addition, the multivariate latent growth model suggested that the initial level of parental monitoring had significant effects, both directly and indirectly through self-control, on the initial level of depression. The linear changes in parental monitoring were associated with the linear changes in self-control; likewise, the linear changes in self-control were related to those in depression. These results imply that parental monitoring and self-control need to be emphasized as a strategy in order to prevent or alleviate adolescents' depression.     

Hindsight,foresight and insight: The challenges of longitudinal qualitative research

This paper describes a UK based longitudinal qualitative study following approximately 100 young people over 9 years. It provides an overview of the research design, methods of data collection and analysis as well as reflecting on the strengths and weaknesses of the strategies employed. We focus on the value of a devolved research design, complementary cross-section and temporal analytic approaches and the effects of being involved in a longitudinal study on researchers and participants. The paper aims to contribute to a growing understanding of the practical, ethical and epistemological challenges presented by longitudinal qualitative methods, and highlights those challenges that are specific to this method as well as those which are amplified by this kind of approach.     

The Effects of Managerial Succession on Organizational Performance

Boyne and Dahya (2002) posit that the means, motives, and opportunitiesavailable to top managers will affect their ability to impactorganizational performance. In this analysis, I test the theoryposited by Boyne and Dahya and expand the model by exploringwhether the performance effects of executive succession differbetween an internal promotion and an external hire. Using Texasschool superintendents as the managers in question, I use pooled,time-series data to test both the immediate and the long-termeffects of managerial succession on performance. The findingsreveal that an immediate, negative effect of executive successionis present only in the case of an externally hired replacementand that the long-term effect of managerial change on organizationalperformance is positive. These findings suggest that publicmanagerial succession does influence organizational performance.     

MEANINGFUL REGRESSION AND ASSOCIATION MODELS FOR CLUSTERED ORDINAL DATA

Many proposed methods for analyzing clustered ordinal data focus on the regression model and consider the association structure within a cluster as a nuisance. However, the association structure is often of equal interest—for example, temporal association in longitudinal studies and association between responses to similar questions in a survey. We discuss the use, appropriateness, and interpretability of various latent variable and Markov models for the association structure and propose a new structure that exploits the ordinality of the response. The models are illustrated with a study concerning opinions regarding government spending and an analysis of stability and change in teenage marijuana use over time, where we reveal different behavioral patterns for boys and girls through a comprehensive investigation of individual response profiles.     

Latent Variable Modeling of Longitudinal and Multilevel Data

An overview is given of modeling of longitudinal and multilevel data using a latent variable framework. Particular emphasis is placed on growth modeling. A latent variable model is presented for three-level data, where the modeling of the longitudinal part of the data imposes both a covariance and a mean structure. Examples are discussed where repeated observations are made on students sampled within classrooms and schools.     

Men's and Women's Pathways to Adulthood and Their Adolescent Precursors

This study compared men's and women's pathways to adulthood by examining how role transitions in education, work, marriage, and parenthood intersect and form developmental pathways from ages 18-30. The study investigated how sociodemographic factors and adolescent experiences were associated with these pathways. We used latent class analysis to analyze longitudinal data from a gender-balanced panel of 808 contemporary young adults. We found three similar latent pathways for both genders, but men and women differed in the timing of marriage and when they began to live with children and the likelihood of combining both roles. The present study points to the continued, though differential, relevance of marriage and family in the transition to adulthood for men and women.     

Disability,partnership and parenting

Partnerships and parenthood can have important effects on economic, social and psychological well-being. We provide new long-term analysis of how disability affects both parental status and partnerships. Analysis of the new Life Opportunities Survey, which is based on social model approaches, demonstrates that disabled people are more likely than non-disabled people to face disadvantages in terms of family formation. Disabled people are more likely to remain single over time, although there is less evidence for any differences in rates of relationship breakdown for those who enter them. Allied to these conclusions, disabled adults are less likely to form households where there are dependent children. These conclusions are supported by longitudinal results from the British Household Panel Survey.     

Evaluation of HIV risk reduction intervention programs via latent growth model

The latent growth model (LGM) has drawn increasing attention in behavioral studies using longitudinal data. The LGM captures the level and trajectory of behavior change, variation in both the initial status and the trend of behavior change, as well as the time-ordered covariation between the initial status and change. This study demonstrates how the LGM can be applied in the evaluation of intervention programs targeting HIV risk behavior among drug users. Multi-group piecewise latent growth models were fit to longitudinal data with three repeated response measures.     

IS THE TIME-SERIES EVIDENCE ON MINIMUM WAGE EFFECTS CONTAMINATED BY PUBLICATION BIAS?

Existing meta-analysis approaches to testing for publication bias are problematic when applied to time-series studies in economics, because changes in parameters can generate spurious evidence of publication bias. We suggest an alternative test in such contexts, and apply it to time-series studies of the effects of minimum wages on employment. In contrast to recent research by Card and Krueger [1995a], we find that the results of published time-series studies of minimum wage effects are consistent with structural change, and that the null hypothesis of no publication bias is not rejected by the evidence. ( JEL B41, C22, C40, J23)     

Analytic strategies for longitudinal networks with missing data

Missing data are often problematic when analyzing complete longitudinal social network data. We review approaches for accommodating missing data when analyzing longitudinal network data with stochastic actor-based models. One common practice is to restrict analyses to participants observed at most or all time points, to achieve model convergence. We propose and evaluate an alternative, more inclusive approach to sub-setting and analyzing longitudinal network data, using data from a school friendship network observed at four waves (N = 694). Compared to standard practices, our approach retained more information from partially observed participants, generated a more representative analytic sample, and led to less biased model estimates for this case study. The implications and potential applications for longitudinal network analysis are discussed.     

FOUR USEFUL FINITE MIXTURE MODELS FOR REGRESSION ANALYSES OF PANEL DATA WITH A CATEGORICAL DEPENDENT VARIABLE

This paper describes and contrasts two useful ways to employ a latent class variable as a mixture variable in regression analyses of panel data with a categorical dependent variable. One way is to model unobserved heterogeneity in the trajectory, or change in the distribution, of the dependent variable. Two models that accomplish this are the latent trajectory model and latent growth curve model for a categorical dependent variable having ordered categories. Each latent class here represents a distinct trajectory of the dependent variable. The latent trajectory model introduces covariate effects on the composition of latent classes, while the latent growth curve model introduces covariate effects on both the "intercept" and the "slope" of growth in logit, which may vary among latent classes.
The other useful way is to model unobserved heterogeneity in the state dependence of the dependent variable. Two models that accomplish this are introduced for a simultaneous analysis of response probability and response stability, and the latent class variable is employed to distinguish two latent populations that differ in the stability of responses over time. One of them is the switching multinomial logit model with a time-lagged dependent variable as its separation indicator, and the other is the mover-stayer regression model.
By applying these four models to empirical data, this paper demonstrates the usefulness of these models for panel-data analyses. Example programs for specifying these models based on the LEM program are also provided.     

Latent Profiles of Discrimination and Socialization Predicting Ethnic Identity and Well‐Being Among Asian American Adolescents

Ethnic identity is rooted in sociocultural processes, but little is known about how social interactions predict its longitudinal changes. Using data from 154 Asian American adolescents, latent profile analysis derived four typologies based on unfair treatment (i.e., discrimination, model minority stereotyping) and ethnic socialization (i.e., cultural socialization, preparation for bias, promotion of mistrust): Low Cultural Salience, High Cultural Salience with Marginalization, Culturally Prepared with Low Mistrust, and High Mistrust/Discrimination. Few gender or generational differences in profile membership were found. Positive outcomes were linked to adolescents attuned to both positive and negative experiences, Culturally Prepared with Low Mistrust, who reported increases in ethnic belonging and decreases in negative emotions. The implications for identity formation and adjustment are discussed.     

Regression artifacts in time-series and longitudinal data

The problem of “regression artifacts” in causal inference, otherwise known as the problem of error and reliable irrelevant variance in “independent” variables used for matching or regression adjustment, is illustrated first in the time-series data where a treatment is triggered by an extreme measure. The “offset effect” in psychotherapy, Underwood's scalloped learning curve, and potential pseudo-effects in AIDS therapies are used as illustrations. The magnitude of such artifacts is computable if the autocorrelation pattern for various lags is known, and thus could be distinguished from genuine effects. For longitudinal studies in which a population of respondents is repeatedly measured, the problem of anchoring the matching or regression adjustments on a single wave of measurement (usually the first) is illustrated as affected by the proximally autocorrelated nature of such measures. Data from a famous study of the effects of job training are reinterpreted in light of this consideration.     

Madam Senator: Growth in Women's State Senate Representation, 1978–2010

Using a dataset of women state senators from all 50 states (1978–2010) and latent growth curve analysis, this article tests two longitudinal theories of the growth of women's political representation over time. Gender salience theory posits that women increase their political representation when they explicitly campaign on their gender. Political climate theory argues that women fare better electorally during periods when domestic issues predominate as opposed to international issues. Results provide support for gender salience theory, but the evidence is too mixed for political climate theory to provide a plausible explanation for the growth in women's state-level political representation. By political party, results suggest that Democratic women were generally advantaged over Republican women; however, Republican women exclusively benefited in the 1992 and 2010 elections. This article concludes with an assessment of the two longitudinal theories and what they may tell us more broadly about women in politics.