Suggestions For Research Design Research Articles

Estimating Homophily in Social Networks Using Dyadic Predictions

Predictions of node categories are commonly used to estimate homophily and other relational properties in networks. However, little is known about the validity of using predictions for this task. We show that estimating homophily in a network is a problem of predicting categories of dyads (edges) in the graph. Homophily estimates are unbiased when predictions of dyad categories are unbiased. Node-level prediction models, such as the use of names to classify ethnicity or gender, do not generally produce unbiased predictions of dyad categories and therefore produce biased homophily estimates. Bias comes from three sources: sampling bias, correlation between model errors and node degree, and correlation between node-level model errors along dyads. We examine three methods for estimating homophily: predicting node categories, predicting dyad categories, and a hybrid “ego–alter” approach. This analysis indicates that only the dyadic prediction approach is unbiased, whereas the node-level approach produces both high bias and high overall error. We find that node-level classification performance is not a reliable indicator of accuracy for homophily. Although this article focuses on a particular version of homophily, results generalize to heterophilous cases and other dyadic measures. We conclude with suggestions for research design. Code for this article is available at https://github.com/georgeberry/autocorr.

Application of Self-determination Theory to Illness Self-management Interventions: Identifying Mediators.

NIMH requires intervention research to utilize an experimental therapeutic approach, identifying mediators to examine causal mechanisms of change. The authors propose utilizing self-determination theory (SDT) to guide research design for self-management interventions, using Wellness Recovery Action Plan (WRAP) as an exemplar. To delineate the theory's relevance to self-management interventions and recovery outcomes, the authors describe the main constructs of SDT and demonstrate its applicability to WRAP and recovery. Suggestions for research design and measurement strategies are provided.

The Effects of Nonresponse Rates on Group-Level Correlations

This study investigates the effects of group member nonresponse in research that examines relationships between group-level constructs aggregated from the responses of individual group members. Simulation data show that the effects of nonresponse vary greatly depending on levels of between-group variability and within-group variability expressed jointly as intraclass correlation, or ICC(1), values. Even high response rates, for example, can produce significantly attenuated correlations between group-level variables with ICC(1) levels commonly reported in group studies. Conversely, even low response rates can yield accurate group-level correlations when ICC(1) levels are high. These effects of nonresponse are equivalent whether nonresponse derives from a random or biased process. Group size appears to play no important role in moderating the effects of response rate. Suggestions for research design and post hoc diagnosis of findings are discussed.