Heterogeneous heterogeneity by default: Testing categorical moderators in mixed-effects meta-analysis

Categorical moderators are often included in mixed-effects meta-analysis to explain heterogeneity in effect sizes. An assumption in tests of categorical moderator effects is that of a constant between-study variance across all levels of the moderator. Although it rarely receives serious thought, there can be statistical ramifications to upholding this assumption. We propose that researchers should instead default to assuming unequal between-study variances when analysing categorical moderators. To achieve this, we suggest using a mixed-effects location-scale model (MELSM) to allow group-specific estimates for the between-study variance. In two extensive simulation studies, we show that in terms of Type I error and statistical power, little is lost by using the MELSM for moderator tests, but there can be serious costs when an equal variance mixed-effects model (MEM) is used. Most notably, in scenarios with balanced sample sizes or equal between-study variance, the Type I error and power rates are nearly identical between the MEM and the MELSM. On the other hand, with imbalanced sample sizes and unequal variances, the Type I error rate under the MEM can be grossly inflated or overly conservative, whereas the MELSM does comparatively well in controlling the Type I error across the majority of cases. A notable exception where the MELSM did not clearly outperform the MEM was in the case of few studies (e.g., 5). With respect to power, the MELSM had similar or higher power than the MEM in conditions where the latter produced non-inflated Type 1 error rates. Together, our results support the idea that assuming unequal between-study variances is preferred as a default strategy when testing categorical moderators.     

Macro- and microlevel factors related to sense of community: The case of temporary neighborhoods in Israel

This study examines macro and micro factors influencing the development of sense of community (Davidson & Cotter, 1980) in two different populations (immigrants and veterans) in new temporary neighborhoods in Israel. At the macrolevel, the major factors examined were population size, population density, number of dwelling units in the site, urbanity of the area, ethnic heterogeneity, and peripheriality of the region. Three kinds of variables were examined on the microlevel: (a) personal attitudes: evaluation of the dwelling unit and satisfaction with public services; (b) social networks; and (c) sociodemographic characteristics. Data were collected from 242 immigrants from the former USSR and from 60 Israeli veterans, residing in 5 different sites. Two different patterns of predictors of sense of community emerged in the two different samples. In the veteran sample, only one macrolevel variable entered the stepwise analysis equation: the number of dwelling units in the site. In the immigrant sample, three significant microlevel factors entered the equation: evaluation of the dwelling unit, external network, and age. The data thus suggest that the major determinant factors underlying sense of community vary for different groups of people. The paper is based on a study initiated and financed by the Ministry of Housing and Construction, the Unit of Social Policy, Israel.     

Perceived Disagreement and Heterogeneity in Social Networks: Distinct Effects on Political Participation

ABSTRACT

Although the coexistence of conflicting opinions in society is the very core of democracy, people’s tendency to avoid conflict could keep them away from political discussion and participation. On the other hand, being exposed to diverse political views could motivate citizens to participate. We conducted secondary analyses on two 2013 ITANES (Italian National Election Studies) probability samples in order to test the hypotheses that perceived network disagreement (between an individual and her/his discussion partners) and heterogeneity (among discussants holding different political opinions) exert independent and opposite effects on political participation through motivation and knowledge. Results converged in showing that disagreement dampened, while heterogeneity encouraged, political participation (voting, propensity to abstain in future, offline and online activism, and timing of vote decision) by decreasing or increasing, respectively, political interest and, in turn, knowledge.     

Group heterogeneity and tolerance: The moderating role of conservation values

Does diversity promote tolerance? We propose that reactions to group heterogeneity depend on individual value priorities. In three studies we investigate how values moderate the effect of raising the salience of group heterogeneity (versus homogeneity) on tolerance. As hypothesized, the findings indicate that conservation values moderate the effects of the saliency of the group's heterogeneity. People high on conservation values were more tolerant when the group's homogeneity was made salient than when its heterogeneity was made salient. People low on conservation were either insensitive to information on the group's heterogeneity (Studies 1 and 2), or were more tolerant when the group's heterogeneity was made salient than when its homogeneity was made salient (Study 3). Values are also directly associated with tolerance: conservation values (as well as SDO and RWA in Study 3) were negatively associated whereas openness-to-change and self-transcendence values were positively associated with tolerance. These findings indicate the importance of integrating individual and contextual variables in theories of group processes.     

Violent, nonviolent, and substance-related offending over the life course in a cohort of males and females treated for substance misuse as youths

Most studies on adolescent offending heterogeneity are based on general population samples, and few include individuals with substance misuse or look specifically at substance-related offending. It is also unclear how offender subtypes develop after young adulthood or how offending heterogeneity varies between genders. This study aimed to identify subgroups of offending among adolescents with misuse problems and to examine associations with offending in adulthood. The study included 1,992 females and males that consulted a clinic for adolescents with misuse problems between 1968 and 1971. Latent Class Analyses were conducted to identify subgroups based on violent and nonviolent offending before age 20. Participants were then followed until age 50 and reexamined regarding violent, nonviolent, and substance-related crimes. Associations between subgroups before age 20 and subgroups age 21-50 were examined. Before age 20, three subgroups were identified among the females and six among the males. Males were more specialized in their offending and demonstrated higher levels of offending. Results pointed to both stability and decrease of violent and nonviolent offending, and to the emergence of substance-related offending in adulthood in both genders. The connection between substance-related crimes and general delinquency in adulthood among individuals treated for substance misuse suggests that interventions should also address substance misuse for reducing the overall volume of crime. This study also highlights the importance of including females in research on offending heterogeneity.     

Investing Even in Uneven Contests: Effects of Asymmetry on Investment in Contests

Many competitions, such as political campaigns, innovative endeavors, sports, or courting rituals, require investment of nonrefundable resources: One contestant wins the prize for the invested amount, whereas all others lose their investments without receiving compensation. Frequently, contests are asymmetric, because of differing resources or prize valuations. Does asymmetry discourage the weaker contestants from investing? Do stronger contestants take this into account and also invest less? Or does the desire to win abolish these differences? Two experiments explored the effects of asymmetry between the contestants—arising from their resources or prizes—on nonrefundable investments. Subjects repeatedly competed in both symmetric and asymmetric contests, enabling within‐subject comparisons. We observed an effect of asymmetry only when it concerned resources: Subjects overall invested less when their resources were unequal, whereas inequality in the prizes did not influence investments. This pattern suggests that winning itself is motivating. The changes between consecutive investments can be explained by reactions to the previous outcome (win or loss): Across all contests, investment decreased following a win and increased following a loss. Copyright © 2015 John Wiley & Sons, Ltd.     

Symptom patterns in children and adolescents with obsessive-compulsive disorder (OCD)

OBJECTIVE: Obsessive-compulsive disorder (OCD) defined at the diagnostic level encompasses divergent symptoms. The present study examines symptom patterns in OCD in children and adolescents in order to search for the presence of diagnostic heterogeneity. SUBJECTS AND METHODS: Two hundred and thirteen outpatients with primary OCD participated. The patients' and primary caretakers' responses on the Children's Yale-Brown Obsessive-Compulsive Scale were used to generate severity ratings for 61 OCD symptoms, and cluster analysed them. RESULTS: A five cluster solution was judged to be the most clinically and theoretically relevant model. The clusters were named: (1) "Mental Rituals, Touching & Ordering", (2) "Contamination & Cleaning", (3) "Superstitions", (4) "Obsessions/Checking & Confessing" and (5) "Somatic Concerns". The symptom patterns of the groups differed, and three groups were particularly homogeneous. In addition, there were differences in severity, number of symptoms and age. CONCLUSIONS: OCD in children and adolescents is a heterogeneous disorder and discrepancies vis-à-vis adult studies indicate that childhood OCD may have developmentally specific traits. The sub-groups obtained are likely to have implications for the assessment of OCD symptoms in children, their prognosis and outcome in treatment research.     

Addendum to: Heterogeneity Coefficients for Mahalanobis' D as a Multivariate Effect Size

In a previous paper (Del Giudice, 2017 [Heterogeneity coefficients for Mahalanobis' D as a multivariate effect size. Multivariate Behavioral Research, 52, 216–221]), I proposed two heterogeneity coefficients for Mahalanobis' D based on the Gini coefficient, labeled H and EPV. In this addendum I discuss the limitations of the original approach and note that the proposed indices may overestimate heterogeneity under certain conditions. I then describe two revised indices H₂ and EPV₂, and illustrate the difference between the original and revised indices with some real-world data sets.     

Model-based recursive partitioning of extended redundancy analysis with an application to nicotine dependence among US adults

Extended redundancy analysis (ERA) is used to reduce multiple sets of predictors to a smaller number of components and examine the effects of these components on a response variable. In various social and behavioural studies, auxiliary covariates (e.g., gender, ethnicity) can often lead to heterogeneous subgroups of observations, each of which involves distinctive relationships between predictor and response variables. ERA is currently unable to consider such covariate-dependent heterogeneity to examine whether the model parameters vary across subgroups differentiated by covariates. To address this issue, we combine ERA with model-based recursive partitioning in a single framework. This combined method, MOB-ERA, aims to partition observations into heterogeneous subgroups recursively based on a set of covariates while fitting a specified ERA model to data. Upon the completion of the partitioning procedure, one can easily examine the difference in the estimated ERA parameters across covariate-dependent subgroups. Moreover, it produces a tree diagram that aids in visualizing a hierarchy of partitioning covariates, as well as interpreting their interactions. In the analysis of public data concerning nicotine dependence among US adults, the method uncovered heterogeneous subgroups characterized by several sociodemographic covariates, each of which yielded different directional relationships between three predictor sets and nicotine dependence.