Overconfidence in interval estimates: What does expertise buy you? q Craig R.M. McKenzie a, * , Michael J. Liersch b , Ilan Yaniv c a Rady School of Management and Department of Psychology, University of California, San Diego, 9500 Gilman Drive, MC 0553, La Jolla, CA 92093-0553, USA b Department of Psychology, University of California, San Diego, 9500 Gilman Drive, MC 0553, La Jolla, CA 92093-0553, USA c Hebrew University, Jerusalem, Israel Received 14 July 2005 Available online 28 March 2008 Accepted by Robyn Dawes Abstract People’s 90% subjective confidence intervals typically contain the true value about 50% of the time, indicating extreme overconfi- dence. Previous results have been mixed regarding whether experts are as overconfident as novices. Experiment 1 examined interval estimates from information technology (IT) professionals and UC San Diego (UCSD) students about both the IT industry and UCSD. This within-subjects experiment showed that experts and novices were about equally overconfident. Experts reported intervals that had midpoints closer to the true value—which increased hit rate—and that were narrower (i.e., more informative)—which decreased hit rate. The net effect was no change in hit rate and overconfidence. Experiment 2 showed that both experts and novices mistakenly expected experts to be much less overconfident than novices, but they correctly predicted that experts would provide narrower intervals with midpoints closer to the truth. Decisions about whether to consult experts should be based on which aspects of performance are desired. Ó 2008 Elsevier Inc. All rights reserved. Keywords: Overconfidence; Expertise; Interval estimates People often express uncertain values in terms of an interval, such as when they estimate their arrival time (‘‘Between 5:00 and 5:30”), another person’s age (‘‘35 to 40”), or next year’s inflation rate (‘‘3% to 5%”). The accu- racy of such estimates is usually measured in terms of hit rate: How often do the intervals contain the true value? Hit rates are often compared to the degree of confidence reported in the intervals. For example, participants might be asked to report low and high values for the populations of various cities such that they are 90% confident that each resulting interval contains the city’s true population. If people were well calibrated, 90% of their 90% confidence intervals would contain the true value. However, true val- ues typically fall within such intervals between 30% and 60% of the time, indicating extreme overconfidence (e.g., Alpert & Raiffa, 1982; Juslin, Wennerholm, & Ols- son, 1999; Klayman, Soll, Gonza ´lez-Vallejo, & Barlas, 1999; Lichtenstein, Fischhoff, & Phillips, 1982; Soll & Klayman, 2004; Teigen & Jørgensen, 2005; Yaniv & Fos- ter, 1997). 1 0749-5978/$ - see front matter Ó 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.obhdp.2008.02.007 q This research was supported by National Science Foundation Grant SES-0551225 and by Israel Science Foundation Grant 344/05. Some of the results were presented at the 2004 Annual Meeting of the Society for Judgment and Decision Making in Minneapolis, MN. * Corresponding author. Fax: +1 858 534 7190. E-mail address: cmckenzie@ucsd.edu (C.R.M. McKenzie). URL: http://psy.ucsd.edu/mckenzie/ (C.R.M. McKenzie). 1 Although we will usually refer to intervals associated with a particular level of confidence as ‘‘subjective confidence intervals” (or ‘‘confidence intervals” for short), readers should be aware that these intervals are sometimes referred to in the literature as ‘‘credible intervals”, ‘‘uncertainty intervals”, ‘‘probabilistic prediction intervals”, and ‘‘fractile assessments” (Teigen & Jørgensen, 2005). www.elsevier.com/locate/obhdp Available online at www.sciencedirect.com Organizational Behavior and Human Decision Processes 107 (2008) 179–191