Special Session Summary Answering Recall Questions: Implications For Consumer Judgment and Choice

Barbara Bickart, Rutgers University-Camden
[ to cite ]:
Barbara Bickart (1996) ,"Special Session Summary Answering Recall Questions: Implications For Consumer Judgment and Choice", in NA - Advances in Consumer Research Volume 23, eds. Kim P. Corfman and John G. Lynch Jr., Provo, UT : Association for Consumer Research, Pages: 28-29.

Advances in Consumer Research Volume 23, 1996      Pages 28-29



Barbara Bickart, Rutgers University-Camden

Over the past several years, there has been a growing body of research examining how people answer survey questions (see Sudman, Bradburn, and Schwarz 1995 for a review). Within this body of research, there have been two distinct areas of attention. One area focuses on how people recall past experiences in order to construct behavioral frequency reports (e.g. Blair and Burton 1987; Menon 1993), while another focuses on how respondents construct subjective judgments, such as assessments of customer satisfaction or overall brand evaluations (e.g. Bickart 1993). In this session, we attempted to bring together these areas of research by examining the relationship between event recall and subjective judgments in a survey context.

Many consumer surveys require that the respondent recall a specific experience or series of experiences from memory. For example, in order to answer a behavioral frequency question, respondents must recall either specific occurrences of the behavior, a rate-of-occurrence, or some other non-numeric impression of how frequently the behavior occurs (e.g. Menon 1993). Likewise, implicit in an evaluation of customer satisfaction is the notion that consumers will base their judgment on previous experiences with the service provider. Consumers may recall a specific experience or an abstraction based on multiple experiences in order to answer such a question. The three papers in this session all presented experimental research that examined how people use recalled experiences to construct either a behavioral frequency judgment or a subjective assessment of customer satisfaction. In addition, the papers examined characteristics of the behavior itself, the individual, or the judgment context that affect how the recalled experiences were used in forming these judgments.

We began the session by examining how characteristics of a behavior might affect the strategy used to estimate its frequency. Frederick Conrad presented the results of two experiments (conducted with Norman Brown) that examined how the abstractness of a category and the distinctiveness of its members affect the strategies and accuracy for frequency estimates. In the first study, carried out as a telephone survey, two broad classes of estimation strategies were identified based on verbal protocols: those using numerical information (enumerating remembered episodes and applying knowledge about rate of occurrence) and those based on non-numerical information (general impressions and memory assessment). Respondents shifted strategies as the nature of the events in question shifted: low frequency distinctive events were enumerated; high frequency, non-distinct events promoted the use of general impressions; regularly occurring events led respondents to use their knowledge about rate of occurrence. Response times corroborate the interpretation of the protocols. A second study focused on the accuracy of different estimation strategies when the abstractness of the target objects varies. Enumeration led to underestimation regardless of category level and non-numerical strategies led to underestimation for superordinate categories. However, non-numerical strategies led to overestimation for basic level categories. While non-numerical strategies make it hard for subjects to fix an upper bound on their estimates, they do provide fairly accurate order information.

While the Conrad and Brown paper focused on how people recall multiple events to construct a frequency estimate, in the second paper, Barbara Bickart showed how the way in which a specific event is recalled can affect a subsequent satisfaction judgment. In an initial study, students recalled either a positive or a negative consumer experience prior to evaluating their satisfaction with the service provider. When subjects were asked to recall a positive experience, they evaluated the service provider more favorably when asked to describe how (versus why) the experience occurred, while when subjects recalled a negative experience, they evaluated the service provider more favorably when asked to describe why (versus how) the experience occurred. Subjects' reported mood mediated these effects. In a second study, focusing only on negative experiences, the respondent's ability and motivation to think about an experience moderated the effects of the question type manipulation (how versus why). Specifically, the question type manipulation only affected the evaluations of subjects who were high in need for cognition. These results suggest that when a consumer experience is recalled, associated emotions or feelings can also come to mind and can affect subsequent satisfaction judgments.

Michaela Wanke then discussed her research (conducted with Claudia Gerke) examining the factors that determine how accessible information related to specific experiences is likely to be used to construct a subsequent judgment. Her findings suggest that recalled experiences are not always used as a direct input to judgment. Of key importance is the perceived diagnosticity of the recalled information. Using an innovative research design, she was able to manipulate subjects' perceptions of the number of positive or negative experiences recalled (e.g. whether they recalled few or many experiences), while holding constant the actual number of experiences recalled. When subjects believed they had recalled many experiences, assimilation effects occurred, presumably because the accessible information was perceived to be diagnostic to a consumer satisfaction judgment. When subjects believed they had recalled only a few experiences, contrast effects occurred. In this case, the accessible information was perceived to be nondiagnostic. Thus, contextual factors affected the perceived diagnosticity of recalled information, and consequently, the way it is used to form a subsequent judgment.

Eric Johnson served as the discussion leader for the session. Eric began by comparing research on answering survey questions to behavioral decision research on preference judgments. He noted several distinctions between these research contexts. First, while preference judgments focus on "looking forward", survey judgments focus on "looking back". As a consequence, research on preference judgments tends to be stimulus-based and research on answering survey questions tends to be memory-based. Second, research on preference judgments typically uses a comparison to a normative standardCthis is usually not the case in research on the survey judgment process.

The general discussion focused on two issues raised by Eric. First, we discussed the methodological problems associated with identifying the theoretical underpinnings of these kinds of effects, particularly with regard to understanding the relationship between the nature of the memory representation of an experience (for example, its "graininess") and a subsequent recall strategy. New methodological approaches may be required in order to advance theory development in this area.

Further, Eric asked how important these kinds of effects are in the real world of survey measurement. For example, he suggested it would be interesting to examine the connections between context effects on satisfaction judgments and subsequent consumer behavior. Members of the audience raised two related issues. First, a number of people were curious about the extent to which the manipulations in the studies reported here parallel actual survey design, particularly in the satisfaction domain. Second, the research reported here is consistent with the idea that answers to survey questions are often constructed on the spot. Thus, we discussed the implications of this research for evaluating the validity of survey responses. Given these and related findings, we may need to rethink traditional conceptions of response validity.


Bickart, Barbara A. (1993), "Carryover and Backfire Effects in Marketing Research," Journal of Marketing Research, 30, (February), 52-62.

Blair, Edward A. and Scot Burton (1987), "Cognitive Processes Used by Survey Respondents to Answer Behavioral Frequency Questions," Journal of Consumer Research, 14, 280-288.

Menon, Geeta (1993), "The Effects of Accessibility of Information in Memory on Judgments of Behavioral Frequencies," Journal of Consumer Research, 20 (December), 431-440.

Sudman, Seymour, Norman M. Bradburn, and Norbert Schwarz (1996), Thinking About Answers: The Application of Cognitive Processes to Survey Methodology. San Francisco: Jossey-Bass Publishers.