Marko Bachl / Michael Scharkow

Some Suggestions on Dealing with Measurement Error in Linkage Analyses

Linkage analysis is a sophisticated media effect research design that reconstructs the likely exposure to relevant media messages of individual survey respondents by complementing the survey data with a content analysis. The main advantage of a linkage analysis is the use of one or more message exposure variables which combine information about media use and media content. However, both constitutive sources are often measured with error: Survey respondents are not very good at reporting their media use reliably, and coders will often make some errors when classifying the relevant messages. In this article, we will first provide an overview of the prevalence and the consequences of measurement error in both data sources, and then discuss a number of possible remedies. While we provide simple recipes for dealing with measurement error using widely available software packages, we also stress the importance of improved measurement as well as further methodological research on linkage analysis.