Examining the cognitive interview in a double-test paradigm

Abstract
The Cognitive Interview (CI) was compared with a standard interview protocol (based on the National Transportation Safety Board) to interview witnesses of a videotaped traffic accident. Witnesses were tested twice, five minutes after viewing the accident and again two weeks later. The CI elicited approximately 70% more correct facts than did the standard interview, and at equivalent accuracy rates, at both the first interview and the second interview. The double-testing procedure generated novel data patterns that may allow us to identify incorrect recollections; other results lead us to question some legal assumptions about the diagnosticity of inconsistent recollections.