Recall Biases and Cognitive Errors
in Retrospective Self-Reports:
A Call for Momentary Assessments
Amy A. Gorin
Arthur A. Stone
State University of New York at Stony Brook
0ne of the primary means of obtaining information from participants in behavioral science research is through the use of retrospective self-reports. Research participants are often required to recall both qualitative and quantitative details about prior events, symptoms, behaviors, and psychological processes (Babor, J. Brown, & Del Boca, 1990). Retrospective self-reports also play an important role in clinical practice. Patients are commonly asked to provide information about their health histories, previous affective states, cognitions, assessment and treatment. The benefits of retrospective selfreports are numerous: They can be developed quickly, are easy to administer and to complete, tend to have strong face validity (Stone, Shiffman, & DeVries, 1999), and are very economical (Baker & Brandon, 1990).
Until recently, participants' confidence in their own reports was the primary indicator of the accuracy of retrospective recall. As participants are usually quite confident in their own recall, the validity of retrospective self-reports has generally been assumed to be high (Fienberg, Loftus, & Tanur, 1985; Read, Vokey, & Hammersley, 1990). However, research in the field of autobiographical memory questions this assumption. Specifically, a growing body of empirical studies indicates that instead of producing reliable and valid information as once assumed, retrospective self-reports are susceptible to numerous recall biases and cognitive errors (Jobe & Mingay, 1991; Jobe, White, Kelley, Mingay, & Sanchez, 1990; Thompson, Skowronski, Larsen, & Betz, 1996). According to Ross (1989, p. 342), retrospective recall appears to be “an active, constructive, schema-guided process, ” which at times is inaccurate. This is not to suggest that retrospective self-reports are' meaningless-they undoubtedly provide valuable information about participants' general perceptions of past events and, perhaps, about their future actions. Rather, it appears necessary to recognize their limitations as tools for gathering accurate, detailed, autobiographical facts and to explore alternative means of obtaining such information.
Given researchers' and clinicians' heavy reliance on retrospective self-reports, this chapter has three goals. First, it reviews autobiographical memory processes involved in retrospective self-reporting such as personal memories and generic personal memory. Second, it provides a summary of the types of recall biases and cognitive errors that impact retrospective self-reports. Rather than reviewing all relevant studies in detail, one or two examples of each recall bias are used to highlight the sorts of errors that can occur. Third, it introduces a relatively new alternative to retrospective self-reports, namely, intensive momentary-based assessments, and discusses instances when this approach may be preferable to traditional self-report methods of collecting information.