Resource sharing in technologically defined social networks

Technologically enabled sharing-economy networks are changing the way humans trade and collaborate. Here, using a novel ‘Wi-Fi sharing’ game, we explored determinants of human sharing strategy. Subjects (N = 1,950) participated in a networked game in which they could choose how to allocate a limited, but personally not usable, resource (representing unused Wi-Fi bandwidth) to immediate network neighbors. We first embedded N = 600 subjects into 30 networks, experimentally manipulating the range over which subjects could connect. We find that denser networks decrease any wealth inequality, but that this effect saturates. Individuals’ benefit is shaped by their network position, with having many partners who in turn have few partners being especially beneficial. We propose a new, simplified “sharing centrality” metric for quantifying this. Further experiments (N = 1,200) confirm the robustness of the effect of network structure on sharing behavior. Our findings suggest the possibility of interventions to help more evenly distribute shared resources over networks.


Recruitment procedure
A total of 1,950 unique subjects participated in our incentivized economic game experiments. Subjects were recruited using Amazon Mechanical Turk (AMT). AMT is an online labor market in which employers contract with workers to complete short tasks for relatively small amounts of money. Workers often receive a baseline payment, plus an additional bonus depending on their performance. Thus, incentivized experiments are easy to conduct using AMT. We used two baseline payments; the first corresponds to the traditional show-up fee ($2.00) and the second to the completion of the game ($2.00).
This latter was necessary for deterring participants from dropping of the game due to its prolonged duration (about 40 minutes on average). Moreover, the participants received a bonus payment, which was proportional to the aggregate Internet bandwidth they received from all their neighbors throughout the game. The exchange rate of the bonus is $1.00 = 200 units.
Issues exist when running experiments online that do not arise in the traditional laboratory. For example, running experiments online naturally implies some loss of control, since the workers cannot be directly monitored as in the traditional lab; experimenters cannot be certain that each observation is the result of a single person (as opposed to multiple people making joint decisions at the same computer), or that one person does not participate multiple times (although AMT goes to great lengths to try to prevent this, and, based on IP address monitoring, it seems to happen very infrequently); and the sample of subjects in AMT experiments is restricted to people who participate in online labor markets (although most physical lab studies are restricted to college undergraduates, who are also far from representative).

Experimental setup
Our participants interacted anonymously over the Internet using customized software playable in a browser window ("Breadboard"; available at http://breadboard.yale.edu ).
We prohibited subjects from participating in more than one session of the experiment by using unique identifications for each subject on AMT. The experiments were conducted from March to September 2016. We conducted 10 sessions for each treatment, and each session comprised of 15 rounds where the players were making repeated decisions.

Instructions and tutorial
Below are screenshots for the initial description of the tutorial and the confirmation tests.
We also show some sample screenshots of the real game. Details are provided as captions after each figure.
Screenshot 1. The main precautions are presented to the players.
Screenshot 2. The first Tutorial screen shows a representation of the "real-world" analogy for the game, in order to increase the realism in the experiment.
Screenshot 3. The 2nd Tutorial screen emphasizes that different houses might have different neighbors and different Internet connections. This helps to mitigate preconceptions of players about their neighbors having identical parameter values.
Screenshot 4. The 3rd Tutorial screen explains that the game is played in rounds; and clarifies how the bonus of each player is calculated. We put emphasis on this latter (by elaborating and using bold typeface fonts) since we need the players to comprehend the factors that affect their monetary compensation.
Screenshot 5. The 4th Tutorial screen initiates the practice session.
Screenshot 6. During the practice session, the players are allowed to select how much Internet they share with their neighbors. Note that we additionally provide detailed instructions and explanations about the game rules.
Screenshot 7. The practice session follows precisely the steps of the actual game, while it also provides additional information and explanation (right panel) to advance the players' comprehension.
Screenshot 8. The practice session continues for another day. This is necessary for emphasizing the time dimension to the players, and the fact that their actions in one round might have impact on their game in the next round.
Screenshot 9. The decision of the 2 nd round (day 2) are shown with an animation.
Screenshot 10. The 9th Tutorial screen concludes the practice session and gives additional guidelines, emphasizing the need for the player to make allocation decisions with the time-window of 90 seconds.
Screenshot 11. The 10th Tutorial screen explains to the players that must answer the comprehension questions, and stresses that they will be able to proceed only if answer correctly in all of them.
Screenshot 12. The first question asks the players what determines their bonus (correct answer is A2). This is very important as we need the players to be aware the payment rules (as this might affect their strategy). Note that for each player and each session the correct answer is allocated in different slot (i.e., not necessarily placed at A2).
Screenshot 13. The second question asks the players about the Internet sharing rule (correct answer is A2 in this display). Clearly, players who have not understood the rules dictating how much Internet (and how) they can allocate to their neighbors should be excluded from the game.