This paper reports three experiments with triadic or dyadic designs. The experiments include the moonlighting game in which first-mover actions can elicit positively or negatively reciprocal reactions from second movers. First movers can be motivated by trust in positive reciprocity or fear of negative reciprocity, in addition to unconditional other-regarding preferences. Second movers can be motivated by unconditional other-regarding preferences as well as positive or negative reciprocity. The experimental designs include control treatments that discriminate among actions with alternative motivations. Data from our three experiments and a fourth one are used to explore methodological questions, including the effects on behavioral hypothesis tests of within-subjects vs. across-subjects designs, single-blind vs. double-blind payoffs, random vs. dictator first-mover control treatments, and strategy responses vs. sequential play.

The recombinant estimation technique of Mullin and Reiley (2006) can be a useful tool for analyzing data from normal-form games. The recombinant estimator falls within a general category of statistics known as U-statistics. This classification has both theoretical and practical implications: (1) the recombinant estimator is optimal (minimum variance) among unbiased estimators, (2) there is a computationally simple method for computing its asymptotic standard error, and (3) the estimation technique can be extended to multiple outcomes and to other types of inferential procedures commonly used for experimental data, such as the sign test. Simulation evidence suggests that researchers should use the asymptotic standard error rather than the standard error of Mullin and Reiley (2006) since the latter exhibits a downward bias.

We study the effectiveness of costly rewards in mitigating excess extraction in a standard Common Pool Resource (CPR) game experiment. We implement two treatments. In the first, rewards are a pure transfer from one player to the other. In the second, the benefits of receiving a reward are higher than the cost of providing it. Referring to the latter as “net positive” rewards, we observe that these are used more frequently than transfer rewards, and that, unlike transfer rewards, they are effective in sustaining cooperation in the CPR game.

In this paper, we present experimental evidence on the effect adverse selection has on coverage choices and pricing in corporate insurance markets. Two sets of experimental data, each generated by experiments utilizing a specific parameterization of a corporate insurance decision, are presented to gauge these effects. In the first, subject behavior conforms to a unique equilibrium in which high risk firms choose higher coverage and contracts are priced accordingly. Insurers act competitively and convergence to equilibrium behavior is marked. In the second set, there is little evidence that subject behavior is consistent with either of the two equilibrium outcomes supported by the experimental setting—pooling by fully insuring losses and pooling by self insuring.

This paper presents a large-scale experiment on the Approval Voting rule that took place during the 2002 French presidential election. We describe the experiment and its main results. The findings are as follows: (i) Such an experiment is feasible, and very well accepted by voters. (ii) The principle of approval voting is easily understood and accepted. (iii) Within the observed political context, compared to the official first-round vote, approval voting modifies the overall ranking of candidates. (iv) The candidates Le Pen and Chirac, more than the others, were able to convert approval votes into official first-round votes.