Steffen Lepa

Combining Inverse Propensity Score Weighting with Media Repertoire Clustering as a Way to Increase Validity and Representativeness Of Longitudinal Cross-media Use Studies

The article provides an outlook on possible methodological answers to the challenge of panel attrition in longitudinal studies of media use. Initially, common prevention strategies, as well as different statistical approaches to deal with full observational unit-nonresponse and their respective limitations are discussed. Then, Inverse Propensity Score Weighting (IPW) is presented as an alternative approach not very common in media and communication research yet. However, as the article argues, IPW exhibits several benefits in terms of easiness, applicability, and complexity reduction compared to usual statistical correction approaches for unit-nonresponse bias. Its practical employment is then demonstrated in detail by depicting the specific reweighting methods employed in a media repertoire study on audio media use conducted by the author. Finally, the article evaluates and discusses IPW’s efficacy, prevailing limitations and possible future use in media and communication research.