Academic journal article Memory & Cognition

The Effect of Category Focus at Encoding on Category Frequency Estimation Strategies

Academic journal article Memory & Cognition

The Effect of Category Focus at Encoding on Category Frequency Estimation Strategies

Article excerpt

We investigated whether category focus at encoding affects how people estimate category frequencies. Participants in three experiments viewed items of various categories. They estimated category frequencies after categorizing them into relevant versus irrelevant categories (Experiments 1-2) or after categorizing versus memorizing them (Experiment 3). Verbal protocols (Experiments 2A and 2B), response latencies (Experiments 2A and 2B), frequency estimate changes (Experiment 2B), and the relationships between objective and estimated category frequencies and instance recall (Experiments 1-3) showed that the participants mainly used availability to estimate category frequencies after memorizing instances (Experiment 3) or after categorizing them into irrelevant categories (Experiments 1-2). After categorizing items into relevant categories, the participants relied more often on stored category frequency information (Experiments 1-3).

People often need to estimate event or category frequencies. For instance, they may wonder how many parking tickets they have had lately or how many mammals they have seen in a certain zoo. In the present research, we investigated how people estimate category frequencies and, more precisely, whether a focus on categorical information while instances are encoded determines how people estimate category frequencies.

Category frequencies or category (set) sizes describe how many instances of categories have occurred or how many instances of certain categories that have occurred fall into a well-defined set. If someone wonders how many mammals he or she has seen in a certain zoo, he or she basically estimates the frequency of the category mammals and, more specifically, mammals in zoo X. Category frequencies differ from event frequencies in that the latter refer to the number of times a stimulus has occurred (e.g., Betsch, Siebler, Marz, Hormuth, & Dickenberger, 1999; Hanson & Hirst, 1988; Manis, Shedler, Jonides, & Nelson, 1993). When estimating the number of parking tickets one has had, one is estimating an event frequency.

Clearly, people can estimate event frequencies (e.g., Brown, 1995, 1997) or behavioral frequencies (Conrad, Brown, & Cashman, 1998; Menon, 1993) in various manners. Tversky and Kahneman (1973) proposed that people estimate category frequencies by using instance availability. One can distinguish between availability by ease and availability by number. Availability by ease refers to the estimation of frequencies on the basis of how easy it is to retrieve (or generate) instances of the category. Availability by number refers to frequency estimation on the basis of how many instances can be retrieved. As such, it implies counting retrieved instances and extrapolating this count to a frequency estimate (see Watkins & LeCompte, 1991).

Ease of retrieval and the number of retrieved instances are usually strongly related (but see, e.g., Schwarz et al., 1991). In fact, ease of retrieval has often been operationalized as the number of instances retrieved (e.g., Curt & Zechmeister, 1984; Lewandowsky & Smith, 1983; Tversky & Kahneman, 1973; Williams & Durso, 1986). Manis et al. (1993) even argued that it is the best operationalization. Therefore, estimating category frequencies through availability is supposed to create a strong correlation between the estimates and the number of instances recalled (e.g., Betsch et al., 1999; Bruce, Hockley, & Craik, 1991; Manis et al., 1993; Watkins & LeCompte, 1991). Moreover, the correlation should reflect a direct, rather than an indirect, relationship (e.g., Betsch et al., 1999; Bruce etal., 1991; Maley, Hunt, & Parr, 2000; Manis et al., 1993).

Many studies support the idea that people estimate category frequencies on the basis of availability. In their classic famous people experiment, Tversky and Kahneman (1973) read a list of 39 men's and women's names to their participants. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.