Between-method convergent validity of four data collection methods in quantitative Means-End-Chain research

Abstract

This paper examines the between-method convergent validity of four alternative data collection methods used in quantitative Means-End-Chain research. The benchmark method is the conventional APT method (i.e. a paper-and-pencil method) in which respondents are asked to indicate existing linkages between product attributes and consumer benefits (i.e. AB), and between consumer benefits and consumer values (BV). The alternative data collection methods are a verbal interview (VI), and two types of computer interviews (CP and CR), each of which differ only in terms of the order in which the AB- and BV-linkages are offered to the respondent. The results show that between-method convergent validity between all four data collection methods is not established. However, when comparing two alternative data collection methods (in particular: CP and VI) with the conventional APT method, between-method convergent validity is supported to some extent. The only data collection method that produces results (i.e. consumer M-E-Cs) which are clearly distinct from those obtained by the conventional APT method, is the computer interview in which the AB- and BV-linkages are offered in a (partial) random order to the respondent (i.e. the CR method).