Terry E. Duncan

Measuring Causal Attributions: The Revised Causal Dimension Scale (CDSII)

Although attribution theory continues to be a fertile area of social psychological research, much of the extant literature has suffered from questionable measurement of the constructs of interest. This is especially true in the case of assigning causal attributions placement in theorized dimensional space. Russells (1982) Causal Dimension Scale represented an important development toward more precise measurement of causal dimensions; however, it has been criticized on a number of fronts. The present report presents the rationale for and initial psychometric properties of a revised version of the scale, the CDSII. Employing data from four studies, a confirmatory factor analysis is reported examining the goodness of fit of the hypothesized four factor oblique structure to the data. The results are discussed in terms of possible applications of the CDSII and the need for further validity testing.

Social support and efficacy cognitions in exercise adherence: a latent growth curve analysis.

Although a considerable amount of research has established the beneficial effects of social support on health and well-being, relatively little work has focused on identifying those processes by which social support influences various health promoting behaviors. Various lines of research demonstrate converging evidence which suggests that self-efficacy may operate as one possible mediator linking psychosocial influences to positive health functioning associated with regular habitual exercise. The present study examined the relationships among social support, self-efficacy, and the ongoing exercise behavior of sedentary adults. Latent growth modeling techniques were utilized to determine whether self-efficacy served a mediational role in the influence of social support on exercise behaviors. Findings are discussed with reference to the role of cognitive and behavioral mechanisms in the relationship between supportive functions of social networks and improvements in health and well-being associated with regular exercise.

A randomised controlled trial of a community intervention to prevent adolescent tobacco use

OBJECTIVE Experimental evaluation of comprehensive community wide programme to prevent adolescent tobacco use. DESIGN Eight pairs of small Oregon communities (population 1700 to 13u2009500) were randomly assigned to receive a school based prevention programme or the school based programme plus a community programme. Effects were assessed through five annual surveys (time 1–5) of seventh and ninth grade (ages 12–15 years) students. INTERVENTION The community programme included: (a) media advocacy, (b) youth anti-tobacco activities, (c) family communications about tobacco use, and (d) reduction of youth access to tobacco. MAIN OUTCOME MEASURE The prevalence of self reported smoking and smokeless tobacco use in the week before assessment. RESULTS The community programme had significant effects on the prevalence of weekly cigarette use at times 2 and 5 and the effect approached significance at time 4. An effect on the slope of prevalence across time points was evident only when time 2 data points were eliminated from the analysis. The intervention affected the prevalence of smokeless tobacco among grade 9 boys at time 2. There were also significant effects on the slope of alcohol use among ninth graders and the quadratic slope of marijuana for all students. CONCLUSION The results suggest that comprehensive community wide interventions can improve on the preventive effect of school based tobacco prevention programmes and that effective tobacco prevention may prevent other substance use.

Exploring associations in developmental trends of adolescent substance use and risky sexual behavior in a high-risk population.

This study examined associations between the development of adolescent alcohol, cigarette, and marijuana use and risky sexual behavior, over time, using latent growth modeling methodology. Gender differences in the development and relationships between use of substances and risky sexual behavior were also examined. Participants were 257 adolescents (mean age = 15.96 years) assessed at three time points over an 18-month period. The intercepts of marijuana with cigarettes and alcohol, and all three substances with risky sexual behavior were significantly related. Development of the three substances showed similar patterns and development of cigarette use covaried with development of risky sexual behavior. There were no significant differences for boys and girls in these relationships. Results are discussed in relation to the need for greater understanding of nonsexual and sex-related problem behaviors and for analyses examining development and change in these behaviors during adolescence.

Latent Variable Modeling of Longitudinal and Multilevel Substance Use Data

This article demonstrates the use of a general model for latent variable growth analysis which takes into account cluster sampling. Multilevel Latent Growth Modeling (MLGR4) was used to analyze longitudinal and multilevel data for adolescent and parent substance use measured at four annual time points. An associative LGM model was tested for alcohol, marijuana, and cigarette use with a sample of 435 families. Hypotheses concerning the shape of the growth curve and the extent of individual differences in the common trajectory over time were tested. The effects of marital and family status and socio-economic status on family levels of substance use were also examined. Findings are discussed in terms of family-level substance use and similarities in developmental trajectories across substances, and the impact of contextual factors on family levels of substance use and development.

Modeling Incomplete Longitudinal Substance Use Data Using Latent Variable Growth Curve Methodology.

Longitudinal data sets typically suffer from attrition and other forms of missing data. When this common problem occurs, several researchers have demonstrated that correct maximum likelihood estimation with missing data can be obtained under mild assumptions concerning the missing data mechanism. With reasonable substantive theory, a mixture of cross-sectional and longitudinal methods developed within multiple-group structural equation modeling can provide a strong basis for inference about developmental change. Using an approach to the analysis of missing data, the present study investigated developmental trends in adolescent (N = 759) alcohol, marijuana, and cigarette use across a 5-year period using multiple-group latent growth modeling. An associative model revealed that common developmental trends existed for all three substances. Age and gender were included in the model as predictors of initial status and developmental change. Findings discuss the utility of latent variable structural equation modeling techniques and missing data approaches in the study of developmental change.

Progressions of Alcohol, Cigarette, and Marijuana Use in Adolescence

This study examined the progressive relations among adolescent use of alcohol, cigarettes and marijuana using latent growth curve analyses. Specifically, the present study examined three models to determine (1) the effect of prior cigarette use on alcohol use and development and the relationship between change in cigarette use and the development of alcohol use (N = 115), (2) the effect of prior alcohol use on cigarette use and development and the relationship between change in alcohol use and the development of cigarette use (N = 199); and (3) the effect of prior alcohol and cigarette use on marijuana use and development, and the relationship between change in alcohol use and cigarette use and the development, of marijuana use (N = 287). Support was found for the relation between prior levels of substance use and involvement in other substances. Cigarette use, in particular, was particularly important in the subsequent involvement of alcohol and marijuana.

Approaches to Testing Interaction Effects Using Structural Equation Modeling Methodology.

Use of structural equation modeling (SEM) methodology to study interactive relationships among latent variables began with the work of Kenny and Judd (1984) who developed a method of testing interactions involving continuous latent variables by forming products of multiple indicator variables. Until recently, there has been considerable difficulty implementing the method in SEM programs. This article reviews a single indicator approach (Joreskog & Yang, 1996) and multiple indicator approaches (Jaccard & Wan, 1995; Ping, 1996) that simplify Kenny and Judds method. An illustrative application using an empirical example examining the interactive effect of perceptions of competence and perceptions of autonomy on exercise intrinsic motivation is presented. Practical issues surrounding the use of these different approaches are discussed.

Piecewise Growth Mixture Modeling of Adolescent Alcohol Use Data

This article addresses issues of heterogeneity in multiple-stage development as it corresponds to qualitatively different development in alcohol use during adolescence. Using a piecewise growth mixture modeling methodology proposed by Muthén (in press), a 2-piece linear growth model capturing growth trajectories in adolescent alcohol use during the transition from middle school (ages 11 to 13) to high school (ages 14 to 17; N = 81) was examined. It was hypothesized that 2 stages of alcohol use development with varying trajectories would exist in these data, the 1st corresponding to development during middle school (Growth Rate 1), followed by a 2nd stage of continuing growth during high school (Growth Rate 2). Results suggested the tenability of the 2-piece linear development in alcohol use and the emergence of 2 latent classes with individually varying transition points. Class 1 showed linear increases only during high school, whereas Class 2 showed a continued, linear growth throughout the middle and high school years. Findings suggest that the sample population under study is heterogeneous and consists of 2 subpopulations, each defined by its unique growth trajectories and individually varying transitional growth processes. The piecewise growth mixture modeling approach is likely to provide researchers with insightful information regarding qualitative differences in adolescent substance use development as well as a potentially useful modeling technique for intervention studies involving evaluation of program effectiveness.

Latent Growth Modeling of Longitudinal Data: A Finite Growth Mixture Modeling Approach

Recent developments in finite mixture modeling allow for the identification of different developmental processes in distinct but unobserved subgroups within a population. The new approach, described within the general growth mixture modeling framework (Muthen, 2001, in press), extends conventional random coefficient growth models to incorporate a categorical latent trajectory variable representing latent classes or mixtures (i.e., the subgroups in the population whose membership must be inferred from the data). This article provides a didactic example of this new methodology with adolescent alcohol use data, which is shown to consist of a mixture of distinct subgroups, defined by unique growth trajectories and differing predictors and sequelae. The method is discussed as a useful tool for mapping hypotheses of development onto appropriate statistical models.