American Scientist

The Science of Self-Report: Implications for Research and Practice. Arthur A. Stone and Jaylan S. Turkkan, eds. 380 pp. Lawrence Erlbaum Associates, 2000. $79.95.

Few in scientific research are immune to the potential complications of self-report. Medical researchers rely on surveys for patient histories, symptom severity, and duration of and adherence to regimens. For epidemiologists and social scientists, frequency of behavior and histories are essential components of complete data collection. Short of continuous monitoring of subjects, it is quite difficult to ensure accurate results.

The Science of Self-Report: Implications for Research and Practice, assembled from presentations at a 1996 National Institutes of Health conference, considers many of the difficulties of self-report, ranging from the malleability of memory to such sensitive events as abortion and sexual abuse.

Perhaps the most widely accessible paper broaches the wide array of ethical concerns inherent in self-report. The researcher faces what David M. Bersoff and Donald N. Bersoff term "competing obligations," in which respect for a study participant's privacy can preclude treatment. The authors conclude: "Full disclosure of confidentiality limitations is of particular concern in self-report research, and not just that involving children, because it is a methodology that allows for the mass collection of extremely sensitive and personal information, much more so than experimental and naturalistic observational research." They advocate specific confidentiality conditions and assurance of researcher competence, as well as data monitoring and intervention planning.

Several papers identify structural weaknesses in the brain's metaphorical filing cabinets and the self-report implications of those flaws. In the more widely accepted of two competing theories, memories are sorted and encoded into discrete categories, creating a hierarchical filing system of types and subtypes. To accurately answer a survey question about an autobiographical incident, one must be able to complete a series of tasks: Understand the question at hand, recall the relevant information, determine the accuracy of the recalled information and format an answer adhering to survey guidelines.

Each of these tasks is subject to different errors, which can yield faulty or incomplete data. Although each paper related to memory assessment has a different focus, the message remains quite constant throughout and is aptly summarized by Roger Tourangeau: "The fact that we can remember something—even have vivid and detailed memories for it—carries no guarantee that we remember it accurately."

In this mental filing system, events may be sorted any number of ways, including chronologically—in calendar units that range from days to seasons to weekdays and weekends, depending on the nature of the event encoded. Researchers have found that context—social (who attended a party), political (who was president at the time) or otherwise—helps to recall an event more accurately. In his paper, Norman M. Bradburn concludes that the most effective calendar is personally constructed of events that can be compared with one another to recall a specific date.

John F. Kihlstrom and his colleagues explore another example of the highly modifiable nature of memory in a fascinating paper on the influence of emotion on memory recall. The authors consider how emotions—during both encoding and retrieval—can bend and mold many of the memories we consider to be hardened images of the past. "Remembering is inherently a reconstructive process with lots of opportunity for bias and error and one's emotional state must cast a light—or a shadow—over that process."

Saul Shiffman explains the most recent and innovative advance in the self-report of momentary states: ecological momentary assessment (EMA), which requires participants to carry a small computer so they can respond to survey questions asked randomly during the day. EMA offers a more impromptu response than in a conventional interview, but the high cost of EMA technology is a practical limitation.

Many researchers and clinicians may be hesitant to fully acknowledge the difficulties of amassing data through surveys, questionnaires and interviews simply because the alternatives are often so unattractive. The Science of Self-Report finds its value by generating awareness of these complications.— Rebecca Sloan Slotnick