Appropriate use of single-item measures is here to stay

In their article, Bergkvist and Rossiter (Journal of Marketing Research, 44, 175-184, 2007) recommended marketing academics to use single-item instead of multiple-item measures for doubly concrete constructs. This recommendation was based on a study showing that the predictive validity of single-item measures was comparable to that of multiple-item measures. Kamakura (2014) presents three criticisms of Bergkvist and Rossiter's study: (1) The correlations used to evaluate predictive validity are inflated by the presence of common-methods variance in the data, (2) the study used concurrent validity as criterion rather than predictive validity, and (3) the multiple-item measures in the study were not corrected for attenuation. A re-analysis of the data from the original study refutes the claims made by Kamakura (2014). The analysis shows that the common-methods variance in the data was negligible and that predicting delayed measures rather than concurrent measures yielded virtually identical results as in the original study. It is also shown that it is possible to estimate single-item reliabilities and correct single-item measures for attenuation, which makes them as predictively valid as multiple-item measures. Thus, there is no reason to change the conclusions and recommendations made in Bergkvist and Rossiter's (Journal of Marketing Research, 44, 175-184, 2007) article. The present article also shows that Kamakura's (2014) analysis of consumer panel data has limitations which casts doubts upon the conclusions drawn from the analysis results. In addition, there is a discussion of the cost, in terms of research quality, that researchers unnecessarily using multiple-items measures pay.