Outreach at the time of the pandemic: an online public good game with punishment
Lory Barile, Warwick University
Published September 2020
This case study relates to the implementation of an online experiment for a taster session organised for a Sutton Trust’s UK Summer School Programme to highlight the relevance of Economics as a broader discipline rather than a specific topic. Students played a public good game with punishments as described in Fehr and Gächter (2000). The idea was to link the experiment to the concept of altruistic punishment and individuals’ behaviour during the pandemic. The experiment was conducted online using Classex. There was a prize awarded for the group who made the highest number of points at the end of the game though punishments came with a cost. Recall that altruistic punishment relates to individuals’ willingness to punish others to promote pro-social behaviour, even where that punishment is costly to them.
In what follows, I will provide some context for the taster session before discussing the main activity in the live-streamed session, followed by some concluding remarks.
The participants in the taster session were Year 12 students from under-represented groups across the UK. In normal conditions, participants to this event come from all over the UK and stay on campus for a few days to experience what being a student at University is like. They attend taster lectures in Economics (and other disciplines), meet with other students and take part in a number of social events. This year, participants joined the Summer School in an online platform and attended the live sessions in MS Teams. In all, 28 students participated in the taster session.
Prior to the live-streamed session, students were asked to read some material on public goods, and, in the morning of the experiment, they received a short video where I introduced myself and provided the instructions of the game. The time allocated to the session was an hour. Thus, although this is not common practice when conducting experiments with students, the video was provided before starting the experiment to save time on reading the instructions in the live session and use it for a post-experiment activity to facilitate discussion. In the video, I guided participants through the basic structure of the game and described the main decisions they had to make during the game, which I will recap below.
Students enjoyed playing the game and found it super accessible and engaging.
The live-streamed session
When the live-streamed session started, I welcomed the students and, with the help of other colleagues, I trained them to join break-out rooms for a post-experiment activity. Students then completed a short Mentimeter survey where they were asked to describe in three words what economics is and what economists do. They were allowed 2 minutes to provide an answer to each of the questions. The objective of having the pre-experiment survey was to get them thinking about what economics is and what economists do to help them to clarify any misconception about the discipline in a follow up session, which was delivered the day after the experiment. Results will be briefly discussed below.
Following this short questionnaire, students had a quick test round of the experiment to familiarise themselves with the different game screenshots before starting to play it in real time. Students were also recommended to annotate their decisions during the game. A score table was created to this end and uploaded on MS Teams prior to the experiment for them to use as an example. This was done to keep track of their strategy during the experiment and see if the introduction of the punishment changed their behaviour.
The main activity of the live session was a simplified public goods game with punishment modified from the format presented inFehr and Gächter (2000). The game was available in Classex. Below I will describe the basic steps of the game including any modifications that I made.
At the beginning of the experiment, students were informed that they had to interact in groups of 4 participants. The composition of the groups remained the same throughout the entire duration of the experiment though the identity of group members was supposed to remain anonymous. They were also told that each participant faced the same series of choices and that, at the end of the experiment, the group with the highest number of points would have won the game.
The game consists of two different stages. I modified the initial instructions of the game to add more explanations and make the terminology used more accessible for students of this age (e.g., private points replaced tokens, and the multiplication factor used in this experiment was clearly defined for them to understand its impact on their payoffs). In the pre-experiment video, I also added examples to clarify how payoffs were calculated in the background at the end of each round of the game. A possible addition to make the video more interactive might be to ask them to work on an exercise to reflect on how payoffs are calculated.
At the beginning of each round, participants received an endowment of 20 private points. In Stage 1, they had to decide how much, if any, of this endowment to allocate to a common project. The common project generated a common benefit which was evenly allocated to all group members at the end of each round. In the instructions that students received, there was no explicit reference to public goods. However, I made a clear reference to the current Coronavirus disease when I suggested them to think about the common project as being the collective actions to help vulnerable people at the time of the pandemic or to avoid average global temperature rises of 2oC by 2050, for example. Participants played Stage 1 for 5 rounds after which they continued to play the game with a single modification.
In Stage 2, in addition to the ‘contribution decision’, students had to decide whether they wanted to reduce or leave equal the other group members’ earnings after receiving feedback on their contributions. This was possible via the allocation of deduction points (which were meant to represent a punishment on free riders). Punishment was costly to all participants. For each point paid to decrease the income of another group member, the payoff of the other group member was reduced by 3 points, whereas the punisher’s income was reduced by 1 point. Participants played Stage 2 for other 5 rounds after which the experiment ended. In each round, I asked students to make their decisions within 20 seconds to avoid running out of time. Time was monitored with a time counter and a buzzer sound was used to warn participants that time had expired.
After the experiment, students were directed in different break-out rooms with a member of staff or a student ambassador (acting as room facilitators) to have a discussion about the experiment they just played. The room facilitators received a set of questions to structure the discussion. These included questions asking them to reflect on the nature of the game they played, the strategy they adopted during the experiment, the reasons for their contribution (if any) to the common project, the use of deduction points and what they represented, and the relevance of the game for economists. The room facilitators were asked to write the students’ answers in a word file uploaded in each of the break-out rooms that were created. The answers to the questions as well as those of the Mentimeter pre-experiment survey were used to conduct some data analysis that was presented in a Q&A session the following day. During this session, the link between the reading material suggested prior to the live session and the experiment was made clear and with clear reference to the pandemic. With reference to the example made in the video related to the pandemic (see above), I took the opportunity to explain to the students what a public good is, and to think about how their behaviour influenced the provision of the public good in the experiment.
Although most of the students were focused on the idea of winning the game and making the most private points, some others clearly mentioned a ‘moral obligation’ as a possible explanation to contribute to the common project. Also, some students were able to link the deduction points to punishing free riders though it was not always clear to them why they received deduction points from others. Some thought for example that they received deduction points because they were contributing a lot to the common pool. The feedback received by students represented a good way to introduce and discuss the main results of the game e.g., I showed them how punishment opportunities increased the average contribution level by round and, as a consequence of this, deduction points decreased towards the end of the experiment. I related this to the idea of altruistic punishment as described in Fehr and Gächter (2000) and a new public enemy, the Covidiot.
Finally, I presented the results of the pre-experiment Mentimeter survey where I asked the students to describe what economics is and what economists do. Regarding the former, the responses were along the lines of ‘money’ or ‘distribution’ and ‘production’. Some students identified a link with Social Science. This gave me an opportunity to clarify any misunderstanding about the discipline and the profession and to discuss with students the relevance of the game for economists. To this end, I linked the game to four different fields in economics (i.e., micro, public, behavioural and experimental economics) and I explained in turn what each field studies. I also mentioned to them that the game represents an example of experiential learning experience which could be easily replicated in an economics class/lecture.
Students enjoyed playing the game and found it super accessible and engaging. They were also particularly positive about the follow up activity in the break-out rooms which gave them a chance to open up some really interesting discussions afterwards. In general, they easily understood how to play the game thanks also to the practice round arranged before playing the experiment. Members of staff and student ambassadors supporting the event were also very helpful and important not only to manage students’ behaviour and expectations but also to guide students’ discussion in the break-out rooms and moderate their questions in the Q&A sessions.
There were no challenges around the setting of the experiment and the activity organised in the break-out rooms. However, there were some difficulties around the time spent on testing how to join and exit the break-out rooms at the beginning of the experiment. A better way to deal with the problem is to provide some training sessions to students before joining events like this in real time. This would have saved plenty of time at the beginning of the live session to be used for providing feedback at the end of the experiment.
Whilst prior to the live session we tested the experiment with colleagues and students to check possible ways of improvement, we didn’t think about a similar event with the summer school participants. This is something to keep in mind if the activity is to be replicated.
Fehr, E., & Gachter, S. (2000). Cooperation and punishment in public goods experiments. American Economic Review, 90(4), 980-994. DOI: 10.1257/aer.90.4.980