Knowledge that Transforms

Most marketing applications of signal detection theory (SDT) produce an estimate of the respondent's memory accuracy based on exposure to a number of advertisements. Marketing practitioners, however, are usually more interested in the performance of an individual advertisement, or elements of that ad. Moreover, advertising recognition paradigms are typically limited to single observations per respondent. The authors present and compare two alternative methodologies that estimate SDT parameters for such designs by pooling recognition performance across respondents. They present two simulations that explore the most efficient methodology and suggest guidelines for selecting appropriate accuracy indices.