University of Iowa
Department of Psychology
Personality and Social Psychology
C5 Seashore Hall
Iowa City, IA 52242
I am interested in the factors that influence everyday judgment and decision making. For example, people estimate the length of the Mississippi River to be longer when they first compare the length to 5,000 miles as compared to people who first compare the length to 200 miles. The comparison value is irrelevant to the final judgment, but people tend to be influenced by this value. Situations where people integrate irrelevant information into the decisions they make are quite common. The research I do investigates situations where this might happen in an attempt to describe and explain judgment and decision making behavior.
Specifically, my areas of interest include: wishful thinking, comparative optimism (and pessimism), the group size effect, anchoring and adjustment, the dud-alternative effect, perceptions of risk and likelihood.
Windschitl, P.D., Rose, J.P., Stalkfleet, M.T., & Smith, A.R. (2008) Are people excessive or judicious in their egocentrism? A modeling approach to understanding bias and accuracy in people’s optimism within competitive contexts. Journal of Personality and Social Psychology, 95, 253-273. [Abstract] [Full Article]
People are often egocentric when judging their likelihood of success in competitions, leading to overoptimism about winning when circumstances are generally easy for competitors but overpessimism when the circumstances are difficult. Yet, egocentrism might be grounded in a rational tendency to favor highly reliable information (about the self) more so than less reliable information (about others). A general theory of probability, called extended support theory (EST) was used to conceptualize and assess the role of egocentrism and its consequences for the accuracy of people’s optimism in three competitions (Studies 1-3, respectively). Also, instructions were manipulated to test whether people who were urged to avoid egocentrism would show improved or worsened accuracy in their likelihood judgments. Egocentrism was found to have a potentially helpful effect on one form of accuracy, but people generally showed too much egocentrism. Debias instructions improved one form of accuracy but had no impact on another. The advantages of using the EST framework for studying optimism and other types of judgments (e.g., comparative ability judgments) are discussed.
Price, P.C., Smith, A.R., & Lench, H.C. (2006). The effect of target group size on risk judgments and comparative optimism: The more the riskier. Journal of Personality and Social Psychology, 90, 382-398. [Abstract] [Full Article]
In 5 experiments, college students exhibited a group size effect on risk judgments. As the number of individuals in a target group increased, so did participants’ judgments of the risk of the average member of the group for a variety of negative life events. This happened regardless of whether the stimuli consisted of photographs of real peers or stick-figure representations of peers. As a result, the degree to which participants exhibited comparative optimism (i.e., judged themselves to be at lower risk than their peers) also increased as the size of the comparison group increased. These results suggest that the typical comparative optimism effect reported so often in the literature might be, at least in part, a group size effect. Additional results include a group size effect on judgments of the likelihood that the average group member will experience positive and neutral events and a group size effect on perceptual judgments of the heights of stick figures. These latter results, in particular, support the existence of a simple, general cognitive mechanism that integrates stimulus numerosity into quantitative judgments about that stimulus.
Windschitl, P.D., Smith, A.R., Rose, J.P., & Krizan, Z. (under review). The Desirability Bias in Predictions: Going Optimistic Without Leaving Realism. [Abstract]
Does desire for an outcome inflate optimism? Previous experiments have produced mixed results regarding the desirability bias, with the bulk of supportive findings coming from one paradigm—the classic marked-card paradigm in which people make discrete predictions about desirable or undesirable cards being drawn from decks. We introduce a biased-guessing account for the effects from this paradigm, which posits that people are often realistic in their likelihood assessments, but when making a subjectively arbitrary prediction (a guess), they will tend to guess in a desired direction. Five experiments tested the desirability bias within the paradigm but also in extensions of it, in order to establish the validity of the biased-guessing account and to distinguish it from other accounts. In addition to supporting the biased-guessing account, the findings illustrate the critical role of moderators (e.g., type of outcome, type of forecast) for fully understanding and predicting desirability biases.
Smith, A.R., Windschitl, P.D., & Bruchmann, K. (under review). Who adjusts? Knowledgeable participants adjust from externally provided anchors. [Abstract]
Previous research has shown that anchoring effects are unaffected by cognitive load, leading to the assumption that people do not effortfully adjust their estimates from externally provided anchors. However, these results might be characteristic of when participants have little knowledge relevant to anchoring questions. The present experiment tested the effects of cognitive load among participants ranging widely in knowledge about the relevant topic. As expected, estimates made by high-knowledge participants—but not low-knowledge participants—were influenced by cognitive load. The authors discuss how this finding provides new insight into the distinction between self-generated and experimenter provided anchors.
Smith, A.R. & Price, P.C. (under review). Sample size bias in the estimation of means. [Abstract]
This research concerns the hypothesis that intuitive estimates of the arithmetic mean of a sample of numbers tend to increase as a function of the sample size. That is, they reflect a systematic sample size bias. A similar bias has been observed when people judge the average member of a group of people on an inferred quantity (e.g., a disease risk; Price, 2001; Price et al., 2006). Until now, however it has been unclear whether this would be observed when the stimuli are numbers, in which case the quantity need not be inferred and “average” can be precisely defined as the arithmetic mean. In two experiments, participants estimated the arithmetic mean of 12 samples of numbers. In the first experiment, samples of from 5 to 20 numbers were presented simultaneously and participants quickly estimated their mean. In the second experiment, the numbers in each sample were presented sequentially. The results of both experiments confirmed the existence of a systematic sample size bias.
Smith, A.R., Windschitl, P.D., & Rose, J. P. Understanding referent-dependent assessments by comparing biases in probability and comparative judgments [Summary]
Research on social-comparative judgments, comparative-optimism judgments, and more generic probability judgments has revealed a set of biases that appear to be interrelated and might (or might not) be explained by similar accounts. For example, above-average effects observed in comparative judgment and subadditivity effects observed in probability judgments clearly have some structural/conceptual relationships. Both comparative and probability judgments require an assessment of the target and referents, however, no previous studies have directly compared these two types of referent-dependent judgments. The present study tested the role of focalism, unpacking, and other accounts by having participants make either comparative or probability judgments that required them to evaluate a focal item (e.g., a specific high-calorie food) in relation to sets of 1, 4 or 9 referent items (e.g., other high-calorie foods). Initial analyses provide support for the predictions of the focalism account as well as partial support for Support Theory across the two judgment types. We provide an explanation for these findings as well as integrate previous research on both comparative and probability judgments.
Smith, A.R. & Windschitl, P.D. Wishful thinking in polychotomous decisions [Summary]
Wishful thinking (an increase in the perceived likelihood of an event because of a desire for that event to occur) is widely accepted by the general population, yet relatively understudied by the scientific community (for a review of wishful thinking, see Krizan & Windschitl, 2007). The majority of wishful thinking studies have examined dichotomous decision tasks (i.e., tasks where one of two options will occur). However, in real world decision tasks, there are often numerous possible outcomes. In this study, wishful thinking is examined in situations where participants make decisions with more than two possible outcomes.
Smith, A.R., Rose, J.P., & Windschitl, P.D. Anchoring effects in health judgments
Summary.
Smith, A.R. & Windschitl, P.D. Anchoring effects in evaluations: Anchors influence judgment but not choice.
Summary.
Smith, A.R. & Windschitl, P.D. Anchoring effects with unambiguous information: Numeric anchors influence answers to math equations. [Summary]
In the majority of studies on anchoring and adjustment, participants have very little information about the target estimate, and/or the estimate under question is subjective in nature. The goal of this study was to investigate anchoring effects in an area where there is an objectively correct answer and participants are presented with complete information. Specifically, participants were briefly shown a math equation. Participants compared the answer of the equation to an anchor value and then provided an estimate of the answer. Half of the estimates were made while under cognitive load. Robust anchoring effects were found and these effects were not moderated by cognitive load. These results extend anchoring research into a new domain and provide information about the mechanisms that produce anchoring effects.
Smith, A.R., Windschitl, P.D., & Scherer, A. Distorting information to support a desired outcome.
Summary.
Smith, A.R., Windschitl, P.D., & Price, P.C. The Effect of Group Size on Non-Social Judgments. [Summary]
The group size effect has been demonstrated when people make risk judgments about the average member of a group (Price, Smith, & Lench, 2006). For example, the average member of a group of 15 is perceived as more likely to get cancer than the average member of a group of 10. Although the group size effect has only been observed when making judgments about groups of people, it may occur when evaluating non-social items as well. Specifically, the average rating of a group of 10 movies may be perceived as being higher than the average rating of a group of 5 movies.
Smith, A.R. When Large Companies are Better (and Worse) than Small Companies: The Effect of Group Size in Managerial Decisions. [Summary]
Price (2001) found that the average member of a company with 9 workers was perceived to have a higher heart attack risk than than the average member of a company with 4 employees. It is likely that the group size effect will generalize to judgments made about areas other than heart attack risk. Therefore, people should view the output of the average member of a large company to exceed that of the average member of a small company. Similarly, a large company should be perceived to commit more work related accidents per worker than a small company. Therefore, a large company will be evaluated as both better (when judging positive factors) and worse (when judging negative factors) than a small company.
