Preference discovery


Is the assumption that people automatically know their own preferences innocuous? We present an experiment studying the limits of preference discovery. If tastes must be learned through experience, preferences for some goods may never be learned because it is costly to try new things, and thus non-learned preferences may cause welfare loss. We conduct an online experiment in which finite-lived participants have an induced utility function over fictitious goods about whose marginal utilities they have initial guesses. Subjects learn most, but not all, of their preferences eventually. Choice reversals occur, but primarily in early rounds. Subjects slow their sampling of new goods over time, supporting our conjecture that incomplete learning can persist. Incomplete learning is more common for goods that are rare, have low initial value guesses, or appear in choice sets alongside goods that appear attractive. It is also more common for people with lower incomes or shorter lifetimes. More noise in initial value guesses has opposite effects for low-value and high-value goods because it affects the perceived likelihood that the good is worth trying. Over time, subjects develop a pessimistic bias in beliefs about goods’ values, since optimistic errors are more likely to be corrected. Overall, our results show that if people need to learn their preferences through consumption experience, that learning process will cause choice reversals, and even when a person has completed sampling the goods she is willing to try, she may continue to lose welfare because of suboptimal choices that arise from non-learned preferences.

This is a preview of subscription content, log in to check access.

Fig. 1
Fig. 2
Fig. 3


  1. 1.

    We distinguish between learning about objective circumstances and learning about one’s own tastes, which Braga and Starmer (2005) refer to as “institutional learning” and “value learning” respectively. Our focus is on value learning, so we assume the agent knows the objective features of all goods. Institutional learning is best separately modeled, e.g., in experimental consumption models (Kihlstrom et al 1984) or the two-armed bandit problem (Rothschild 1974).

  2. 2.

    Brezzi and Lai (2000) show, in another theoretical study, that learning when facing a multiple-armed bandit will be incomplete, but it is for a different reason (discounting) than what we study.

  3. 3.

    Chuang and Schechter (2015) find, in developing country contexts, very little stability in preferences within a person over years, except in survey measures of self-reported social preferences. However, their interpretation is that the experimental measures they study are not good measures of preferences in these contexts.

  4. 4.

    In a related paper, Delaney et al (2019), we develop a formal theory exploring the extensive margin of preference discovery.

  5. 5.

    Problems with the server caused fatal timeouts for some potential subjects. Of the 606 who did not complete the experiment, 547 (90.3%) had made no choices by the time they stopped. Most of these likely had server timeouts.

  6. 6.

    The post-experiment questionnaire asked a comprehension question that posed a simplified version of the experiment’s choice problem. 82.4% of subjects answered correctly. Including only those who answered correctly produces qualitatively identical results except that the Mann–Whitney test for the effect of noise on efficiency becomes non-significant and the effect of noise on efficiency becomes significant at the 10% level in the Tobit regression for \(T = 10\). This paper reports results from the full sample of subjects.

  7. 7.

    Subjects made other non-maximizing choices as well. Of 103,950 good-round pairs, subjects chose a value between 0 and 1 (less than a meaningful consumption experience) 322 times (or 0.3% of the time), and a value less than 0 a total of 14 times (less than 0.1% of the time). 99.3% of the time, subjects chose an integer between 0 and 6.

  8. 8.

    Choices inconsistent with believed preferences have little impact on our experiment’s results. Excluding subjects who make these non-myopic-maximizing choices yields qualitatively similar results with only a few changes: In Table 3, the difference in efficiency across income levels becomes marginally significant (\(p = 0.090\)), while the difference in efficiency across noise levels ceases to be statistically significant (\(p = 0.118\)). In Table 4, the number of remaining rounds becomes significant at \(p < 0.001\) in the second model, while in the third model, noise becomes marginally significant (\(p = 0.064\)) as does income (\(p = 0.063\)).

  9. 9.

    Recall that the numeraire value is known with certainty and thus is always “learned.” Results are similar excluding the numeraire. We also considered a version of specification III that used individual period dummies and found qualitatively similar results.

  10. 10.

    The calculation of this threshold is shown in the electronic supplementary material. A key assumption is that in all future periods, only the numeraire good and this good will be available. As we show in the online appendix, an agent has achieved maximally experimental full relevant discovery according to this definition if she has tried all goods i with priors at least as large as \(X_i^t \,=\, 64.5 - \sigma + \sqrt{240.25-240\cdot \frac{q_i \cdot y \cdot (T-t)}{1+q_i \cdot y \cdot (T-t)}}\). Note that this threshold moves higher as t progresses.


  1. Aghion, P., Bolton, P., Harris, C., & Jullien, B. (1991). Optimal learning by experimentation. The Review of Economic Studies, 58(4), 621–654.

    Article  Google Scholar 

  2. Andersen, S., Harrison, G. W., Lau, M. I., & Rutstrom, E. E. (2008). Lost in state space: Are preferences stable? International Economic Review, 49(3), 1091–1112.

    Article  Google Scholar 

  3. Ariely, D., Loewenstein, G., & Prelec, D. (2003). “Coherent arbitrariness”: Stable demand curves without stable preferences. The Quarterly Journal of Economics, 118(1), 73–105.

    Article  Google Scholar 

  4. Armantier, O., Lévy-Garboua, L., Owen, C., & Placido, L. (2016). Discovering preferences: A theoretical framework and an experiment.

  5. Becker, G. S. (1996). Accounting for tastes. Harvard: Harvard University Press.

    Google Scholar 

  6. Braga, J., & Starmer, C. (2005). Preference anomalies, preference elicitation and the discovered preference hypothesis. Environmental and Resource Economics, 32(1), 55–89.

    Article  Google Scholar 

  7. Brezzi, M., & Lai, T. L. (2000). Incomplete learning from endogenous data in dynamic allocation. Econometrica, 68(6), 1511–1516.

    Article  Google Scholar 

  8. Chen, D. L., Schonger, M., & Wickens, C. (2016). oTree: An open-source platform for laboratory, online, and field experiments. Journal of Behavioral and Experimental Finance, 9, 88–97.

    Article  Google Scholar 

  9. Chuang, Y., & Schechter, L. (2015). Stability of experimental and survey measures of risk, time, and social preferences: A review and some new results. Journal of Development Economics, 117, 151–170.

    Article  Google Scholar 

  10. Cooke, K. (2017). Preference discovery and experimentation. Theoretical Economics, 12(3), 1307–1348.

    Article  Google Scholar 

  11. Coursey, D. L., Hovis, J. L., & Schulze, W. D. (1987). The disparity between willingness to accept and willingness to pay measures of value. The Quarterly Journal of Economics, 102(3), 679–690.

    Article  Google Scholar 

  12. Cox, J. C., & Grether, D. M. (1996). The preference reversal phenomenon: Response mode, markets and incentives. Economic Theory, 7(3), 381–405.

    Article  Google Scholar 

  13. Dasgupta, U., Gangadharan, L., Maitra, P., & Mani, S. (2017). Searching for preference stability in a state dependent world. Journal of Economic Psychology, 62(Supplement C), 17–32.

    Article  Google Scholar 

  14. Delaney, J. J., Jacobson, S. A., & Moenig, T. P. (2019). A theory of preference discovery.

  15. Easley, D., & Kiefer, N. M. (1988). Controlling a stochastic process with unknown parameters. Econometrica, 56(5), 1045–1064.

    Article  Google Scholar 

  16. Eckel, C. C., El-Gamal, M. A., & Wilson, R. K. (2009). Risk loving after the storm: A bayesian-network study of hurricane katrina evacuees. Journal of Economic Behavior and Organization, 69(2), 110–124.

    Article  Google Scholar 

  17. Kahneman, D., & Snell, J. (1990). Predicting utility. In R. M. Hogarth (Ed.), Insights in decision making: A tribute to Hillel J (pp. 295–310). London: University of Chicago Press.

    Google Scholar 

  18. Kahneman, D., Wakker, P. P., & Sarin, R. (1997). Back to Bentham? explorations of experienced utility. The Quarterly Journal of Economics, 112(2), 375–405.

    Article  Google Scholar 

  19. Keller, G., & Rady, S. (1999). Optimal experimentation in a changing environment. The Review of Economic Studies, 66(3), 475–507.

    Article  Google Scholar 

  20. Kihlstrom, R. E., Mirman, L. J., & Postlewaite, A. (1984). Experimental consumption and the ‘Rothschild Effect’., Studies in Bayesian econometrics, New York; Amsterdam and Oxford: North-Holland; distributed in U.S. and Canada by Elsevier Science, New York (vol. 5, pp. 279–302).

  21. Lichtenstein, S., & Slovic, P. (2006). The construction of preference. Cambridge: Cambridge University Press.

    Google Scholar 

  22. List, J. A. (2003). Does market experience eliminate market anomalies? The Quarterly Journal of Economics, 118(1), 41.

    Article  Google Scholar 

  23. Loewenstein, G., & Adler, D. (1995). A bias in the prediction of tastes. The Economic Journal, 105(431), 929–937.

    Article  Google Scholar 

  24. Noussair, C., Robin, S., & Ruffieux, B. (2004). Revealing consumers’ willingness-to-pay: A comparison of the BDM mechanism and the Vickrey auction. Journal of Economic Psychology, 25(6), 725–741.

    Article  Google Scholar 

  25. Piermont, E., Takeoka, N., & Teper, R. (2016). Learning the krepsian state: Exploration through consumption. Games and Economic Behavior, 100, 69–94.

    Article  Google Scholar 

  26. Plott, C.R. (1996). Rational individual behaviour in markets and social choice processes: The discovered preference hypothesis. In: K.J. Arrow, et al (Eds.) The rational foundations of economic behaviour: Proceedings of the IEA Conference held in Turin, Italy, IEA Conference (vol. 114, pp. 225–250). New York: St. Martin’s Press; London: Macmillan Press in association with the International Economic Association.

  27. Rothschild, M. (1974). A two-armed bandit theory of market pricing. Journal of Economic Theory, 9(2), 185–202.

    Article  Google Scholar 

  28. Scitovsky, T. (1976). The joyless economy: An inquiry into human satisfaction and consumer dissatisfaction. Oxford: Oxford University Press.

    Google Scholar 

  29. Shogren, J. F., Cho, S., Koo, C., List, J., Park, C., Polo, P., et al. (2001). Auction mechanisms and the measurement of WTP and WTA. Resource and Energy Economics, 23(2), 97–109.

    Article  Google Scholar 

  30. Shogren, J. F., Shin, S. Y., Hayes, D. J., & Kliebenstein, J. B. (1994). Resolving differences in willingness to pay and willingness to accept. American Economic Review, 84(1), 255–270.

    Google Scholar 

  31. Thaler, R. H., & Sunstein, C. R. (2008). Nudge: Improving decisions about health, wealth, and happiness. London: Penguin.

    Google Scholar 

  32. van de Kuilen, G., & Wakker, P. P. (2006). Learning in the Allais paradox. Journal of Risk and Uncertainty, 33(3), 155–164.

    Article  Google Scholar 

  33. Weber, R. A. (2003). Learning with no feedback in a competitive guessing game. Games and Economic Behavior, 44(1), 134–144.

    Article  Google Scholar 

  34. Wilson, T. D., & Gilbert, D. T. (2005). Affective forecasting: Knowing what to want. Current Directions in Psychological Science, 14(3), 131–134.

    Article  Google Scholar 

Download references


We are grateful for helpful comments from the editor and two anonymous referees. For advice early on, we thank Yongsheng Xu, Annemie Maertens, and participants at FUR 2012, SABE/IAREP/ICABEEP 2013, and seminars at Williams College and George Mason University, and we particularly thank CeMENT 2014 participants Brit Grosskopf, Muriel Niederle, J. Aislinn Bohren, Angela de Oliveira, Jessica Hoel, and Jian Li for detailed feedback. We gratefully acknowledge funding from the Williams College Hellman Fellows Grant.

Author information



Corresponding author

Correspondence to Sarah Jacobson.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Supplementary material 1 (pdf 179 KB)

Supplementary material 2 (pdf 904 KB)

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Delaney, J., Jacobson, S. & Moenig, T. Preference discovery. Exp Econ 23, 694–715 (2020).

Download citation


  • Discovered preferences
  • Preference stability
  • Learning

JEL Classification

  • D81
  • D83
  • D01
  • D03