ContemporaryApproaches to Assessing Mediation in Communication Research

tionships. Whether it is examining the persuasive impact of public service announcements on attitudes and behavior, determining the impact of viewing political debates on political knowledge or voter turnout, or assessing whether success in achieving one’s Internet browsing goals prompts greater interest in e-commerce, communication scholars frequently conduct research to answer questions about cause. Data analysis usually focuses on examining if the putative causal variable, whether manipulated or measured, is related to the outcome using a linear model such as analysis of variance or linear regression. In many arenas of research, such analyses, when accompanied by good research design, are sufficient to answer the question as to whether variation in X causes variation in Y. But deeper