School choice and information: An experimental study on matching mechanisms

Abstract

We present an experimental study where we analyze three well-known matching mechanisms-the Boston, the Gale-Shapley, and the Top Trading Cycles mechanisms-in different informational settings. Our experimental results are consistent with the theory, suggesting that the TTC mechanism outperforms both the Boston and the Gale-Shapley mechanisms in terms of efficiency and it is slightly more successful than the Gale-Shapley mechanism regarding the proportion of truthful preference revelation, whereas manipulation is stronger under the Boston mechanism. In addition, even though agents are much more likely to revert to truth-telling in lack of information about the others' payoffs-ignorance may be beneficial in this context-the TTC mechanism results less sensitive to the amount of information that participants hold. These results therefore suggest that the use of the TTC mechanism in practice is more desirable than of the others.

Cognitive Ability and Games of School Choice

“…Starting from Chen and Sönmez [2006], a growing literature investigates school choice using laboratory experiments, exploiting the advantage of controlling students' preferences by design [Calsamiglia et al, 2010;Chen et al, 2015;Chen and Kesten, 2012;Klijn et al, 2013;Pais and Pintér, 2008]. In these papers, the authors are mostly interested in comparing truth-telling rates, and identifying the rules of thumb used by applicants under different mechanisms.…”

Self-Selection in School Choice

“…To this end, market designers have advocated the use of strategy-proof mechanisms, which guarantees that students can do no better than submitting preferences truthfully (Roth (2008)). However, non-truthful behaviors are not only possible in theory, as truth-telling is a weakly dominant strategy, but also well documented in lab (Chen and Sönmez (2006) and Pais and Pintér (2008)). …”

Constrained School Choice: An Experimental Study

“…Each subject only knows his own payoff table and his district school (the school for which he has the highest priority). A recent experiment study by Pais and Pintér (2008) shows that, in the unconstrained case, the less information subjects have about others' payoffs the more likely they are to be truthful for BOS, SOSM and TTC. That is, providing subjects with virtually no information about their opponents' payoffs represents the least favorable situation to reject that the mechanisms induce truthful behavior.…”