A large body of research now exists which demonstrates that the cognitive interview (CI) increases the reporting of correct recall from various population groups. It is now necessary to determine from both a theoretical and an applied perspective which of the CI techniques, or combination of techniques are largely responsible for this CI superiority effect. The two aims of this study were (1) to examine the relative effectiveness of each of the four original CI mnemonics and (2) to assess their efficacy as a function of age. A total of 125 participants; 34 first-year undergraduate students, 44 children aged 8–9 years and 47 children aged 5–6 years viewed a video-recording of an accident. Forty-eight hours later each participant was individually interviewed and randomly assigned to one of six instruction groups; (1) context reinstatement, (2) change perspective, (3) change order, (4) report everything, (5) report everything + context reinstatement combination (RE + CR), or (6) a control instruction to ‘try again’. Each of the individual CI mnemonics were found to be of equal benefit and to occasion no more recall than the ‘try again’ control. However, the RE + CR combination resulted in significantly more correct recall compared to the individual mnemonics. There was no effect of age group on the efficacy of the various instructions. Copyright © 2002 John Wiley & Sons, Ltd.