Experimental Design Details
This experiment consists of several stages, and participants are informed that the instructions for each stage are to be given as they go along. Subjects start with instructions to the main task in which they need to decide how much of their endowment to allocate as input in production. Subjects are informed that this task consists of 5 sequences of decision making, where each sequence consists of 10 periods. To avoid the complexity of economic terms such as "utility" and "habit stock", the utility function is presented as output and cost of input and explained in general terms. While providing the interpretation and equations for output, cost of production, and return, the subjects are presented with a graphical illustration of those functions. Further, subjects can test how their output, cost, and returns in a period varies with the opening stock using a slider.
Instructions carefully explained that at the end of the experiment, one sequence (out of 5 sequences) is to be randomly selected by the computer, and the converted monetary value of accumulated returns in that sequence is added to their payoff. During the instructions, the subjects are presented with a series of questions to check their understanding of the task where they would not be able to proceed to the next page without correctly answering those questions. Further, an example interface of the main task is presented with the instructions on how to make the decision using the slider, how the subjects can see their current return and accumulated return up to the current period, and relevant explanation regarding the information provided in that interface.
After the instructions, a "Trial Sequence" ( one sequence of the main decision making task with similar periods) is provided to the subjects for experimenting. The information before commencement and after completion of the trial sequence emphasized that the subjects' decisions during the trial sequence do not affect their final payoff. After the trial sequence, the subjects start sequence one, which may contribute to the final payoff (if it is randomly selected for the payment). Before starting the first period of each sequence, subjects are reminded that that sequence may be selected for the payment. At the end of a sequence, subjects receive information on their outcome of that sequence and the converted monetary value that would be added to the subject's payoff if that sequence is selected for payment. Then, subjects are asked to elicit their confidence in their decisions at that sequence. More explicitly, the subjects are asked to estimate their earnings in the last sequence as a percentage of the maximum possible earnings. If their estimation is within the 5% range of actual realized earnings as a percentage of maximum possible and if that sequence is selected for the payment, subjects earn 10 points in addition to their total return in that sequence. Note that subjects receive complete information about their earnings during a sequence. Therefore when estimating their earnings as a percentage, subjects are actually guessing the maximum possible earnings. The instructions and procedure until the end of the first sequence are similar across all the subjects in control groups and treatment groups.
Treatments:
1. Other's decisions in the selected sequence (DecisionOnly)
2. Other's decisions and accumulated returns in the selected sequence (DecisionReturn)
3. Other's decisions and confidence in the selected sequence (DecisionConfid)
From sequence two onward, the treatments are allocated. Treatments are randomly allocated, and throughout the rest of the sequences, subjects receive the same treatment which is allocated at the beginning. That is, this study uses a between-subject treatment design. Further, regardless of the treatment, all subjects are reminded of their decision, outcomes, and confidence in the previous sequence to avoid biases due to selective memory recall. The control group receive no more information than their own decisions, outcomes, and confidence in the previous sequence. In addition, the treatment groups receive the peer's information as described above. Note that this information is selected from subjects who performed the task in an earlier session in control groups.
The selection of information from subjects in previous sessions controls for any reflection issues. That is, since the information is selected from a subject who performed the task earlier, they do not get influenced by the subjects who observe that information. Moreover, all subjects in a specific treatment receive the same information across all the sessions. This controls for the endogeneity of information provided.
After the main sequences of decision making, subjects face two attempts of "Race game", a risk elicitation task, and a demographic survey. "Race game" is a task where players need to reach the given target number (in this experiment, the target number is 12) first. A player can add one or two to the number selected by the previous player in each round, where players take alternate turns. Subjects play against the computer, and the computer is always the player one. This game has a strictly dominant strategy that can be obtained by backward induction (with three steps). Since the computer is always the first player, subjects have the competitive advantage of winning the game. If a subject misses the dominant strategy at any stage in this game, they lose the game. If a subject wins, 25 points are added to the payoff. Further, to identify the risk appetite of the subjects, this study uses the task described in Gneezy and Potters (1997) with 20 points as endowment and 2.5 as the multiplier where the probability of success is 0.5. Finally, subjects need to answer a short demographic survey. After subjects complete the survey, they receive information on their earnings in each task and total final earnings.