Performance of a reciprocity model in predicting a positive reciprocity decision

This economic experiment initiates in evaluating a model's performance in predicting a decision. The reciprocity model is measured its accuracy rate in prediction and informativeness as aspects of the model's performance. Seventy-nine undergraduate students voluntarily joined the experiment. They made decisions contingently in designed situations as the first player in a dictator game and all roles in trust-share games. The study controls effects of choice set (equal split, competitive, and different social welfare choices) and framing effect. The result shows that the model has high performance in both prediction and informative. Furthermore, it shows an existence of the loss aversion behavior, and a significant relationship between decisions in the dictator game and the trustshare games. The study suggests that the more complicated model may not be marginally useful in predicting decision in the positive reciprocity situations.