What is the longitudinal evidence for causal intergroup contact effects? - A comparative multi-method re-analysis of 21 published studies

One important and increasingly used approach to estimate causal effects of intergroup contact research are longitudinal studies. One of the main challenges in longitudinal research is to control for all relevant (un)observed confounders to generate unbiased estimates of causal contact-attitude effects. The commonly used cross-lagged panel model has been criticized for insufficiently accounting for unobserved confounders, and improved statistical models have been proposed, including the full-forward and random-intercept cross-lagged panel model. To explore the extent of potential bias between the different models, we investigated the robustness of longitudinal contact-attitude effects in a systematic re-analysis using 31 datasets from 21 published studies (N = 33,348). We found substantial discrepancies showing notably smaller, non-significant longitudinal contact-attitude effects in models with increased control for unobserved confounders. We discuss the theoretical and methodological implications of these results and the design and analytical requirements for robust and informative future longitudinal (contact) research.