USING RESPONSE LATENCY TO DETECT INACCURATE RESPONSES IN A COMPUTERIZED LIFE-STYLE ASSESSMENT

This study evaluated the usefulness of response latency to differentiate between fake good and honest self-report to an adapted version of the Computerized Lifestyle Assessment (CLA). Seventy clients from an addiction treatment center were randomly assigned to either a fake good then honest or an honest then fake good instruction group. Multivariate analysis revealed that content scores under fake good instructions were greater than those for honest self-report. Differences in latencies for fake good and honest self-report were largely due to instruction order. Shorter latencies were associated with the second testing condition, regardless of instructions. However, latency scores for Emotional Health were shorter for fake good than honest self-report for both instruction groups. Discriminant analysis indicated that latency scores for Emotional Health provided the greatest discriminant power in differentiating between fake good and honest responses. The use of response latency is discussed in terms of passive and active computerized assessment strategies. © 1990.