Experimental Design
The online experiment consisted of three parts. Part one of the experiment began by explaining the nature of the incentives during the program. Participants were told that an anonymous benefactor with an interest in promoting healthy living and supporting the broader community provided funds to incentivize YOTA members to attend their local YMCA more frequently. During the month of the Grow & Thrive program, a $2 donation was made on participants' behalf for each day a participant visited the YMCA, up to a total donation of $60 per person (e.g., 30 visits).
In part two, participants were told that they might also be randomly selected to participate in SRP. We explained that if a participant was selected into this additional program, he/she would receive an email at the end of Grow & Thrive, which would: (1) list the names of everyone in SRP; (2) list their attendance during Grow & Thrive; and (3) list the total donations generated by them during Grow & Thrive. We explained that only participants in SRP would receive and be listed in the email.
We then elicited people's willingness to pay for receiving (or avoiding) social recognition (e.g., being part or not of the recognition program) using a combination of the strategy method and the Becker-DeGroot-Marschak elicitation method (BDM). The strategy-proof method contained 11 questions asking participants to state whether at the end of the program they wanted to participate in the social recognition program for different numbers of realized attendances during Grow & Thrive, and then eliciting for each question how much they were willing to pay (between $0 and $8) to guarantee that their choice was implemented (the BDM component). The categories of possible visits were the following: 0 visits, 1 visits, 2 visits, 3 visits, 4 visits, 5 or 6 visits, 7 or 8 visits, 9 to 12 visits, 13 to 17 visits, 18 to 22 visits, and 23 or more visits.
Each of the eleven questions had the following structure: “If I go X times to the YMCA during Grow & Thrive I would prefer to participate (NOT participate) in the personal recognition program". Participants were then asked to state, for each of the 11 levels of possible attendance, how much of the experimental budget of $8 they would be willing to give up to guarantee that their decision about social recognition was implemented. The wording was “You said you would rather be part (NOT be part) of the personal recognition program if you go X times to the YMCA. How much of the $8 reward would you give up to guarantee that you will indeed be part (NOT be part) of the personal recognition program?" Although the procedure was involved, we told participants that, first and foremost, it was in their best interest to answer truthfully. The details were then explained in simple and plain language.
To preserve random assignment, as well as to minimize any negative inferences that could be drawn about those not in the social recognition group, we guaranteed that the BDM responses would be used to determine assignment only with 10% chance. We explained to participants that they would have a 10% chance of receiving an additional $8 reward in the form of an Amazon gift card, the contents of which would be used to “pay” to implement their choices. If they were selected to be in that 10%, at the end of the experiment a computer would check how many times they actually attended the YMCA. The computer would then randomly choose a number between $0 and $8. If the computer chose a value smaller or equal to the amount they would give up to receive (or avoid) social recognition for the level of their actual attendance during the experiment, then their favorite decision about the SRP would be implemented. In this case, their extra reward would either be equal to $8 with 50% chance, or equal to $8 minus the amount chosen by the computer with 50% chance. If instead the computer chose a value larger than the amount that they would give up to receive/avoid social recognition, then their favorite decision would be implemented with only 50% probability and they would receive a reward of $8. Participants were also told that each draw for each participant is independent.
From subjects' perspective, this procedure is equivalent to a second price sealed-bid auction against an unknown bidder. Note that providing a 50-50 chance to receive the desired option whenever random draws are above participants' bids was necessary, as otherwise participants would have had an incentive to misrepresent their true preferences in the first part of the “yes/no” elicitation and then simply bid zero.
Because others' behavior plays a crucial role in social recognition payoffs it was important to ensure that participants had accurate beliefs about others' behavior. Prior to making their decisions about being part of the SRP, participants were therefore informed about the average attendance to their local YMCA during the prior month.
Part three elicited participants' beliefs about their future attendance during Grow & Thrive. We asked how many times they believed they would attend during the month of the experiment if: (1) they happened to be randomly selected to participate in the social recognition program; (2) they happened not to be selected to SRP; (3) if they happened not to be part of Grow & Thrive (and consequently of SRP). Finally, we reminded participants that a computer would randomly determine whether they would be part of SRP, and we asked them to explicitly agree to participate in Grow & Thrive.