A General Class of Nonparametric Models for Ordinal Categorical Data

This paper presents a general class of models for ordinal categorical data that can be specified by means of linear and/or log-linear equality and/or inequality restrictions on the (conditional) probabilities of a multiway contingency table. Some special cases are models with ordered local odds ratios, models with ordered cumulative response probabilities, order-restricted row association and column association models, and models for stochastically ordered marginal distributions. A simple unidimensional Newton algorithm is proposed for obtaining the restricted maximum-likelihood estimates. In situations in which there is some kind of missing data, this algorithm can be implemented in the M step of an EM algorithm. Computation of p-values of testing statistics is performed by means of parametric bootstrapping.     

Does the Association between Religiosity and Substance Use Vary between Adolescence and Early Adulthood?

The present study examines whether the association between substance use and two dimensions of religiosity—institutional and personal—changes as individuals transition from adolescence to early adulthood. Using data from the National Longitudinal Study of Adolescent to Adult (Add) Health, logistic regression models with interaction terms were run to test the hypothesis that the association between an individual's religiosity and his or her level of substance use is stronger during early adulthood compared to adolescence. This hypothesis is based on the observation that the shared family environment largely explains the association between religiosity and substance use during adolescence, while individual‐level characteristics largely explain this association during early adulthood. Overall, my findings support the research hypothesis, but only for the association between institutional religiosity and stigmatized types of substance use (smoking and marijuana use). This study highlights the fact that studies exploring the association between religiosity and substance use should consider potential variation between life stages.     

Log-Multiplicative Association Models As Latent Variable Models For Nominal and/or Ordinal Data

Associations between multiple discrete measures are often due to collapsing over other variables. When the variables collapsed over are unobserved and continuous, log-multiplicative association models, including log-linear models with linear-by-linear interactions for ordinal categorical data and extensions of Goodman's (1979, 1985) RC(M) association model for multiple nominal and/or ordinal categorical variables, can be used to study the relationship between the observed discrete variables and the unobserved continuous ones, and to study the unobserved variables. The derivation and use of log-multiplicative association models as latent variable models for discrete variables are presented in this paper. The models are based on graphical models for discrete and continuous variables where the variables follow a conditional Gaussian distribution. The models have many desirable properties, including having schematic or graphical representations of the system of observed and unobserved variables, the log-multiplicative models can be read from the graphs, and estimates of the means, variances, and covariances of the latent variables given values on the observed variables are a function of the log-multiplicative model parameters. To illustrate some of the advantageous aspects of these models, two examples are presented. In one example, responses to items from the General Social Survey (Davis and Smith 1996) are modeled, and in the other example, panel data from two groups (Coleman 1964) are analyzed.     

Models and methods for characterizing the structural parameters of groups

This paper outlines and illustrates the application of loglinear models for characterizing the structural parameters of groups, as conceptualized by Blau (1974, 1977). Models for nominal parameters are those in which an ingroup outgroup distinction is embodied. Models for graduated parameters introduce distance effects. These models appear to be useful supplements to spatial techniques for identifying patterns of differential association, and they have implications for the calculation of proximity estimators in the use of those techniques.     

Leadership in Voluntary Associations: The Case of the “International Association of Women”

Three models of leadership in voluntary associations have been proposed in the literature: democratic leadership, oligarchy, and leadership by default. Through an intensive case study of leadership structure, differences in the attitudes of members and leaders at three hierarchical levels, and differences between the attitudes and behaviors of aspirants and nonaspirants to leadership in a women's service association, this article examines the degree of fit between these models and a specific organization. Data is drawn from questionnaires, annual reports, and interviews. The results fail to conform to any of the existing models, suggesting instead a fourth model, leadership for self-development—in which leaders are motivated primarily by a desire to develop administrative and interpersonal skills.     

Statistical Methods and Graphical Displays For Analyzing How The Association Between Two Qualitative Variables Differs Among Countries, Among Groups, Or Over TimePart II: Some Exploratory Techniques, Simple Models, and Simple Examples

We introduce some simple models that can be viewed as special cases of the general model and general approach to analyzing the association in multiway tables that we first presented in 1998. We use two empirical examples to demonstrate the flexibility and utility of the simple models introduced here and of our general approach.     

History of Childhood Abuse and Intergenerational Support to Mothers in Adulthood

The authors examined the association between maternal childhood abuse and the frequency of providing social support to mothers when the victim is an adult and investigated the mediating effects of internal working models (self‐esteem, positive relations with others) and maladaptive coping. Using data from the National Survey of Midlife Development in the United States II, they analyzed a sample of 887 adult children using a structural equation modeling approach. The results showed that maternal childhood abuse was associated with providing less frequent emotional support to mothers, although the authors did not find significant mediating effects of internal working models and maladaptive coping in the association between child abuse and intergenerational support giving. The findings suggest implications for theory, practice, and policy to address specific concerns and possible challenges that adults with a history of childhood abuse may experience in the relationship with their mothers.     

Antecedents of rural adolescent alcohol use: a risk factor approach.

The present study examines the association between risk factors and alcohol use for a sample of young adolescents in a rural eastern community. Family relations, family structure, marks in school, participation in academic activities, frequency of church attendance, and deviant behavior were found to be significantly associated with alcohol use two years later. No gender or age differences were found in these predictors of alcohol use. These six risk variables were combined to form a risk index. A 3 x 2 x 2 (User group by Gender by Grade) ANOVA was used to examine the association between the risk index score at Year 1 and level of alcohol use at Year 3. Only the main effect for User group was significant. Thus, the number of risk factors at Year 1 was predictive of alcohol use at Year 3. The risk index also predicted frequency of alcohol use in a replication sample. Implications for the risk factor approach, prevention, and intervention are discussed.     

Racial Discrimination, Post Traumatic Stress, and Gambling Problems among Urban Aboriginal Adults in Canada

Little is known about risk factors for problem gambling (PG) within the rapidly growing urban Aboriginal population in North America. Racial discrimination may be an important risk factor for PG given documented associations between racism and other forms of addictive behaviour. This study examined associations between racial discrimination and problem gambling among urban Aboriginal adults, and the extent to which this link was mediated by post traumatic stress. Data were collected via in-person surveys with a community-based sample of Aboriginal adults living in a mid-sized city in western Canada (N = 381) in 2010. Results indicate more than 80 % of respondents experienced discrimination due to Aboriginal race in the past year, with the majority reporting high levels of racism in that time period. Past year racial discrimination was a risk factor for 12-month problem gambling, gambling to escape, and post traumatic stress disorder (PTSD) symptoms in bootstrapped regression models adjusted for confounders and other forms of social trauma. Elevated PTSD symptoms among those experiencing high levels of racism partially explained the association between racism and the use of gambling to escape in statistical models. These findings are the first to suggest racial discrimination may be an important social determinant of problem gambling for Aboriginal peoples. Gambling may be a coping response that some Aboriginal adults use to escape the negative emotions associated with racist experiences. Results support the development of policies to reduce racism directed at Aboriginal peoples in urban areas, and enhanced services to help Aboriginal peoples cope with racist events.     

Random-Effects Modeling of Categorical Response Data

In many applications observations have some type of clustering, with observations within clusters tending to be correlated. A common instance of this occurs when each subject in the sample undergoes repeated measurement, in which case a cluster consists of the set of observations for the subject. One approach to modeling clustered data introduces cluster-level random effects into the model. The use of random effects in linear models for normal responses is well established. By contrast, random effects have only recently seen much use in models for categorical data. This chapter surveys a variety of potential social science applications of random effects modeling of categorical data. Applications discussed include repeated measurement for binary or ordinal responses, shrinkage to improve multiparameter estimation of a set of proportions or rates, multivariate latent variable modeling, hierarchically structured modeling, and cluster sampling. The models discussed belong to the class of generalized linear mixed models (GLMMs), an extension of ordinary linear models that permits nonnormal response variables and both fixed and random effects in the predictor term. The models are GLMMs for either binomial or Poisson response variables, although we also present extensions to multicategory (nominal or ordinal) responses. We also summarize some of the technical issues of model-fitting that complicate the fitting of GLMMs even with existing software.     

The Role of Family Income in the Intergenerational Association of AFDC Receipt

Daughters from low‐income families who did not receive Aid to Families with Dependent Children (AFDC) are compared to daughters from low‐income families who did receive assistance to better account for family income in the intergenerational association of AFDC. The research models the timing of a daughter's first birth and, for those who become mothers, the number of years a daughter receives any AFDC. Women whose families received AFDC during their childhood are more likely to receive AFDC as adults relative to those women whose families did not receive AFDC. Controlling for family income in a more careful and comprehensive way than past research explains part but not all of the effects of childhood AFDC receipt. Among daughters from chronically poor families, however, parental AFDC use is not associated with additional years of AFDC participation. Although the present research cannot confirm or deny a causal role for parental welfare use, the intergenerational transmission of financial resources does explain part of the intergenerational association of welfare use.     

Statistical Methods and Graphical Displays for Analyzing How the Association Between Two Qualitative Variables Differs Among Countries, Among Groups, Or Over Time: A Modified Regression-Type Approach

Many interesting sociological questions pertain to how the association between two variables depends on a third variable. In sociological applications, the third variable often pertains to countries, to subgroups of a population, or to time periods. We propose a regression-type approach that specifies that the log-odds- ratios that describe the two-way association of interest are a linear function of latent scores for the third variable. Additive and multiplicative models currently in use by researchers are special cases of the regression-type model. To illustrate the utility of the regression-type approach, we apply this approach to analyze (1) data on occupational mobility in the United States, Britain, and Japan (comparing mobility in these countries) and (2) data on the association between religion and voting behavior in U.S. presidential elections from 1968 to 1992 (comparing this association in the different elections). We also introduce here graphical displays that can be used to obtain worthwhile information about goodness-of-fit and to aid in substantive interpretation.     

The Relationship Between Use of Sexually Explicit Media and Sexual Risk Behavior in Men Who Have Sex with Men: Exploring the Mediating Effects of Sexual Self-Esteem and Condom Use Self-Efficacy

ABSTRACT

Objectives: This study tests the following 3 hypotheses: (1) there is a direct association between consumption of sexually explicit media (SEM) depicting non-condom use and sexually transmitted infection (STI)-related sexual risk behavior among men who have sex with men (MSM); (2) the association between SEM consumption and STI-related sexual risk behavior is mediated by men's sexual self-esteem; and (3) the relationship between SEM consumption and sexual risk behavior is mediated by condom use self-efficacy. Methods: A cross-sectional, Internet-based survey on exposure to SEM and sexual behavior of 1,391 MSM in the United States was conducted in 2011. Results: The results confirmed Hypotheses 1 and 3, while Hypothesis 2 was rejected. Accordingly, a significant association between the use of SEM picturing condom use and STI-related sexual risk behavior among MSM was found. Likewise, we found that the association between the use of SEM and sexual risk behavior was mediated by condom use self-efficacy in an indirect path. However, SEM did not influence sexual risk behavior via sexual self-esteem. Conclusions: To promote STI prevention, the actors in SEM may be used as role models in managing condom use in sexual contexts.     

Multipartnered Fertility and Depression among Fragile Families

We used data from the Fragile Families and Child Wellbeing Study to examine the association between multipartnered fertility (MPF)-when parents have children with more than one partner-and depression. Random effects models suggested that MPF is associated with a greater likelihood of depression, net of family structure and other covariates. However, these associations disappeared in more conservative fixed effects models that estimated changes in MPF as a function of changes in depression. Results also suggested that social selection may account for the link between MPF and depression for fathers (but not mothers), as depressed fathers with no MPF were more likely to have a child by a new partner four years later. Ultimately, MPF and depression may be reciprocally related and part of broader processes of social disadvantage.     

An Ecological Approach to Understanding Youth Violence: The Mediating Role of Substance Use

The authors tested an ecological model that posits mediating variables (substance use and mental health) in the association between ecological factors (family closeness, school closeness, and peer closeness) and youth violence in a sample of 4,783 adolescents. Models including substance use present significantly less total effect between ecological factors and youth violence than do models without substance use. Additional probing of significant mediation effect using the Sobel test was performed and suggested that substance use did function as a mediator in the hypothesized path. Considerations of adolescent violence should recognize the possible role of ecological factors and how their influence may vary by substance use.     

An Ecological Approach to Understanding Youth Violence: The Mediating Role of Substance Use

The authors tested an ecological model that posits mediating variables (substance use and mental health) in the association between ecological factors (family closeness, school closeness, and peer closeness) and youth violence in a sample of 4,783 adolescents. Model including substance use present significantly less total effect between ecological factors and youth violence than do models without substance use. Additional probing of significant mediation effect using the Sobel test was performed and suggested that substance use did function as a mediator in the hypothesized path. Considerations of adolescent violence should recognize the possible role of ecological factors and how their influence may vary by substance use.     

Different pathways explain alcohol-related problems in female and male college students

ABSTRACT

Objectives: Comprehensive models elucidating the intricate associations of depressive symptoms, coping motives, alcohol use, alcohol-related problems (ARPs), and gender among young adults have been scarcely examined. This study investigated relationships among these variables and the effect of gender on these pathways. Methods: College students (N = 163; 49.7% female) completed self-report measures on alcohol consumption, depressive symptoms, coping motives, and ARPs. Results: Structural equation modeling showed that the association between depressive symptoms and ARPs was mediated by coping motives in both females and males. However, frequency of heavy alcohol use mediated the association between depressive symptoms and ARPs in females but not in males. Conclusions: Different models explain the association between depressive symptoms and ARPs in male and female college students. Prevention programs aimed at reducing ARPs should focus on increasing alcohol screening among students with depressive symptoms, teaching coping skills, and emphasizing moderation in alcohol consumption.     

Parents' Relationship Quality and Children's Behavior in Stable Married and Cohabiting Families

Although an extensive literature has shown that family structure is linked with child well‐being, less well understood is how the dynamics within families affect children, in particular the extent to which positive mother–father relationship quality is linked with children's outcomes. In this study the authors used data from the Fragile Families and Child Wellbeing Study (N = 773) to examine how couple supportiveness in stable coresident families is related to children's externalizing and internalizing behavioral problems over ages 3 through 9. Using latent growth curve and fixed effects models, they found that parents' greater supportiveness has a slight association with lower levels of children's behavioral problems. Using cross‐lagged structural equation models to examine the direction of the association, they also found some evidence that parents' relationship quality and children's behavioral problems are reciprocally related. Overall, this study suggests that more positive couple interactions are beneficial for children residing with both of their biological parents.     

Cultures of Melancholia in Late Capitalism—A Reflection Response to Discussion by Jeanne Wolff Bernstein

Our reply to Jeanne Wolff Bernstein claims that she is an ideal reader for the type of analysis we offered: because she is familiar with the European literary tradition, because she so obviously harbors a sense of the interface of literature and psychoanalysis, and—not least—because she herself has so meticulously studied the nuanced details of our diagnosis. We concur, furthermore, with the melancholia diagnosis she attributes to us and understand why our analysis led her—almost by association—to remember the models of earlier theorists of melancholy and depression. However, we caution that these three are merely a few among many models and that they should not, in any case, be considered definitive (as if models could ever be ultimate arbiters in an enterprise such as the posthumous analysis of a historical figure). Finally, in our reply we petition for a more historical approach than the one found in Bernstein's response or in the writings of the three thinkers whose approach has influenced her thinking.     

EFFECT DISPLAYS FOR MULTINOMIAL AND PROPORTIONAL-ODDS LOGIT MODELS

An "effect display" is a graphical or tabular summary of a statistical model based on high-order terms in the model. Effect displays have previously been defined by Fox (1987, 2003) for generalized linear models (including linear models). Such displays are especially compelling for complicated models—for example, those including interactions or polynomial terms. This paper extends effect displays to models commonly used for polytomous categorical response variables: the multinomial logit model and the proportional-odds logit model. Determining point estimates of effects for these models is a straightforward extension of results for the generalized linear model. Estimating sampling variation for effects on the probability scale in the multinomial and proportional-odds logit models is more challenging, however, and we use the delta method to derive approximate standard errors. Finally, we provide software for effect displays in the R statistical computing environment.