Contamination of responses due to extreme and midpoint response style can confound the interpretation of scores, threatening the validity of inferences made from survey responses. This study incorporated person-level covariates in the multidimensional item response tree model to explain heterogeneity in response style. We include an empirical…

Steinberg and Monahan's (2011) reanalysis of the Teen Media longitudinal survey of adolescents does not meet prevailing standards for propensity score analysis and therefore does not undermine the original conclusions of the Brown, L'Engle, Pardun, Guo, Kenneavy, and Jackson (2006) analysis. The media do matter in the sexual socialization of…