Abstract
Although the issue of method bias is a well-known threat to the validity of self-reports, it has seldom been addressed in empirical media effects studies. This may be problematic because independent and dependent variables are often measured using similar methods. Thus, reported correlations may not only reflect common variance between the constructs of interest but also indicate common variance between the measurement methods. In this article, we present two cultivation studies investigating to what extent first-order cultivation relationships are influenced by common method variance. In both studies, controlling for a theoretically unrelated marker variable (Study 1: the estimated prevalence of bakers; Study 2: the estimated prevalence of bricklayers) diminished correlations between television-viewing frequency and prevalence estimates. This finding, referred to as the bricklayer effect, suggests a more thorough discussion on common method variance in cultivation research is needed.
Additional information
Notes on contributors
Lennert Coenen
Lennert Coenen (M.Sc., KU Leuven, 2013) is a Ph.D. candidate at the Leuven School for Mass Communication Research (KU Leuven). His research is mainly directed at theory development in the area of media effects on public opinion.
Jan Van den Bulck
Jan Van den Bulck (Ph.D., KU Leuven, 1996) is a full professor at the Leuven School for Mass Communication Research (KU Leuven). His research interests include the role of the media in the social construction of reality, cognitive approaches to media effects, and fiction effects (e.g., cultivation theory).