Lab in the Field Experiments and Collective Action Research: Evidence from a study of Ugandan producer organizations

Lab in the field experiments that incorporate behavioral games into socially meaningful settings are an interesting addition to the social sciences’ research tool-kit. This article describes how I used lab in the field experiments to study the mechanisms through which small groups can overcome collective action problems. Namely, I took behavioral games out of the aseptic walls of the laboratory and brought them to the field, conducting research with members of pre-existing groups – i.e., Ugandan producer organizations – that face collective action problems on a regular basis. By adopting an innovative research design that combines behavioral games and observational data, I was able to isolate some of the mechanisms that make group members cooperate in real life.

In very general terms, behavioral games (BGs hereafter) are abstract situations in which individuals have to allocate resources between themselves and other players, and they are uniquely suited to capture actors’ interdependence in decision-making. BGs have been originally developed to study general traits of human behavior, such as altruism, cooperation, and sanctioning (Marwell and Ames, 1979; Camerer, 2003; Fehr and Gachter, 2002; Fehr and Gintis, 2007) using convenience samples (often college students), and experimental settings and protocols that guaranteed complete anonymity. Over the last decade, however, BGs have been deployed by some scholars in order to capture macro-cultural variations across societies (Henrich et al., 2004; Henrich et al., 2010; Herrmann et al., 2008), and, more recently, they have also been used to measure individual or group differences that stem from contextual variations and individual experiences (Fearon et al., 2009; Barr, 2003; Karlan, 2005; Ermisch and Gambetta, 2010). This last development is critical: the fact that behavioral games have been shown to be sufficiently sensitive to detect differences between individuals within a society makes them an exceptionally powerful tool for research in those fields that rely on “hard-to-measure” concepts, such as altruism, trust, and reciprocity, among others.

In collaboration with Guy Grossman, I conducted a study of Uganda’s largest recent rural development intervention – the Agriculture Productivity Enhancement Project (APEP) –, whose goal is to support small farmers integration into commercial farming by exploiting economies of scale and increasing farmers’ productivity. To achieve their objectives, these producer organizations have to overcome classic problems of collective action that affect some of their core activities, e.g., collective marketing and quality improvement. In order to understand how these farmer groups solve collective action problems, our research design was conceived in order to capture both centralized solutions, in which a leader or small elite are the locus of coordination, control and enforcement of cooperative efforts, and decentralized solutions, in which cooperation emerges through repeated interaction and social networks and mechanisms of group solidarity, reciprocity, and peer pressure are at work.

In particular, one of our working hypotheses was that the legitimacy of the organization leaders and their willingness to monitor and sanction non-cooperative behavior greatly influences group outcomes. Unfortunately, this hypothesis is hard to test relying exclusively on observational data: although we do show, relying on observational data, that there exists a positive relationship between the perceived legitimacy of the organization managers and members’ level of cooperation, this result cannot be deemed conclusive, because of selection and measurement issues, and the possibility of reverse causality.

Thus, to demonstrate that mechanisms of centralized sanctioning and leadership legitimacy positively affect cooperation we designed a novel adaptation of the public goods game (PGG), which is a game that captures how players balance the trade-off between individual and group interest (Camerer 2003).1 Experimental evidence shows that, in PGGs participants initially contribute, on average, between 40 and 60% of their endowment. In repeated games, however, conditional cooperators who wish to avoid being exploited by free-riders gradually refrain from cooperation, leading to a drop in contributions in subsequent rounds (Fischbacher et al., 2001; Ostrom, 2000). By contrast, when participants are allowed to punish other subjects, overall levels of contribution increase, since conditional cooperators can discipline defectors (Fehr and Gachter, 2002).

In our adaptation of the PGG, members of the farmer organizations were randomly assigned to three different conditions: a baseline condition (with no monitoring authority), a random monitor condition, in which the monitor was selected through a random lottery, and an elected monitor condition in which the monitor was elected by the players using a secret ballot. Through this experimental component, we were able to show that in the presence of a centralized sanctioning system, individuals are more likely to cooperate. Moreover, we also show that the process of monitor selection is consequential: elected monitors are perceived as more legitimate and thus elicit greater contributions to the public good (Baldassarri and Grossman, 2011).

Performing laboratory experiments in a field setting and with members of pre-existing groups that face collective action problems on a regular basis does not only increase the external and ecological validity of the findings; more importantly, it makes it possible to relate experimental behavior to ‘real life’ behavior. In the third part of the analysis we compare organization members’ behavior in behavioral games with their level of cooperation in the producer organization and show that farmers’ deference to authority in the controlled setting predicts cooperation in their natural environment: those individuals who contribute more in the elected monitor condition are also the more cooperative members of the farmer organizations, thus suggesting that centralized sanctioning and leader legitimacy are relevant factors in explaining organizational outcomes (Grossman and Baldassarri, 2012).

Does this mean that we have demonstrated the existence of a causal relationship between perceived legitimacy of the leader, and cooperation? Within the boundaries of the experimental setting, we can confidently conclude in favor of such causal relationship. This result alone, however, would be discarded by some social scientists for its modest external and ecological validity. The value added of carrying out the behavioral experiment in a field setting is the capacity to test whether the behavior observed in the ‘real’ setting is consistent with the mechanisms captured in the controlled experimental setting. In this analytical framework, the lab in the field experiment is therefore used as a ‘petri dish’ in order to isolate the mechanisms that are likely to be at work in ‘real life’.

I used a similar logic in a second set of lab in the field experiments, in which I explored the mechanisms through which horizontal social relations may have contributed to enhance cooperation among farmer group members (Baldassarri, 2014). Repeated interaction and social networks are commonly considered viable solutions to collective action problems. However, assessing the relationship between social networks and pro-social behavior is not sufficient in order to determine the building blocks of cooperation. To understand how cooperation emerges in a specific social setting, it is important to focus on how patterns of social relations affect actors’ motives and their expectations about others’ behavior. Although previous scholarship has identified few mechanisms that may trigger cooperation among interconnected actors, these mechanisms have been rarely compared with each other, or tested across diverse settings.

To address this issue, I identified and systematically measured four general mechanisms, i.e., generalized altruism, group solidarity, reciprocity, and the threat of sanctioning, using different variants of the dictator and public goods game. Namely, I used the basic version of the dictator game,2 in which deciders have to divide an endowment between themselves and a stranger, to measure generalized altruism, and a version of the DG in which the recipient is a farmer group member to measure group solidarity. Second, I used different variants of a Public Goods Game (PGG) to study the relative effectiveness of reciprocity via communication and the threat of sanctioning.3 Namely, I randomly assigned participants to one of three variants of a iterated PGG: a baseline condition, a sanctioning condition with an elected monitor, and a communication condition in which participants were allowed to discuss their strategy.

Following the logic of the previous field experiment, first I established a positive relationship between position in the social network structure and propensity to cooperate in the producer organization relying on observational data, and then used farmers’ behavior in dictator and public goods games to test for the different mechanisms that may account for such relationship. Results show that cooperation in ‘real-life’ is not induced by other-regarding preferences like altruism or group solidarity. Rather, repeated interaction through communication favors the development of mechanisms of reciprocity (Baldassarri, 2014).

Although, drawing from my own experience, I have here focused on their use for the study of collective action, lab in the field experiments have been recently deployed in diverse settings and to study a variety of social and political phenomena, ranging from micro-level studies of social preferences to the evaluation of institutional performance (for some examples in political science, see Grossman 2011). All social sciences can greatly benefit from this new and exciting research tool in their quest for empirically rigorous explanations.


* Delia Baldassarri is Associate Professor of Sociology at the Department of Sociology, New York University, New York.

1 In a classic PGG, participants anonymously decide how to split an initial endowment between private and public accounts. What players put in the private account remains theirs; what is contributed to the public account is doubled and redistributed evenly among all group members, regardless of their personal contribution. The most profitable outcome for the group occurs when all players contribute their entire endowments. Nonetheless, the most profitable strategy for the individual is to keep the entire endowment in his private account and benefit from what everyone else contributes to the public account.

2 In a classic DG two subjects are given a common endowment. One of them, the decider, has to decide how to divide the money between him/herself and the other player, the receiver. The decider keeps whatever s/he has decided to allocate to him/herself, while the receiver takes home whatever s/he has been given. The DG is conducted under conditions of anonymity. If deciders were completely selfish, they would keep the entire endowment to themselves. Instead, individuals typically share between 20 and 30% of their endowments. A few share up to a half, whereas the modal behavior is to give nothing. Behavior in DG is usually interpreted as an expression of other-regarding preferences (Camerer, 2003).

3 While sanctioning is widely considered the most common solution to collective action problems, scholars have also shown that face-to-face communication produces substantial increases in cooperation (Ostrom et al., 1992).


  • Baldassarri, D. (2014), ‘Cooperative Networks: Group Solidarity, Reciprocity, and Sanctioning in Ugandan Farmer Organizations’, 33, American Journal of Sociology.
  • Baldassarri, D. and Grossman, G., (2011), ‘Centralized Sanctioning and Legitimate Authority Promote Cooperation in Humans’, Proceedings of the National Academy of Sciences, 108(27): 11023-11027.
  • Barr, A. (2003), ‘Trust and Expected Trustworthiness: Experimental Evidence from Zimbabwean Villages’, The Economic Journal
  • Camerer, C. (2003), Behavioral game theory: experiments in strategic interaction. New York, N.Y.: Russell Sage Foundation.
  • Ermisch, J. and Gambetta D. (2010), ‘Do strong family ties inhibit trust?’, Journal of Economic Behavior & Organization, 75: 365 – 376.
  • Fehr, E. and Gintis, H. (2007), ‘Human Motivation and Social Cooperation: Experimental and Analytical Foundations’, Annual Review of Sociology, 33: 43–64.
  • Fehr, E., and Gächter, S. (2002), ‘Altruistic Punishment in Humans’, Nature, 415(6868): 137–40.
  • Fischbacher, U., Gächter, S. and Fehr, E. (2001), ‘Are People Conditionally Cooperative? Evidence from a Public Goods Experiment’, Economics Letters, 71: 397–404.
  • Grossman, G. (2011), ‘Lab-in-the-field Experiments’, Newsletter of the APSA Experimental Section, 2(2): 13-19.
  • Grossman, G. and Baldassarri , D. (2012), ‘The Impact of Elections on Cooperation: Evidence from a Lab in the Field Experiment in Uganda’, American Journal of Political Science, 56(4): 964-985.
  • Karlan, D. S. (2005), ‘Using Experimental Economics to Measure Social Capital and Predict Financial Decisions’, American Economic Review, 95: 1688–1699.
  • Habyarimana, J., Humphreys, M., Posner, D. N., and Weinstein, J. M. (2009), Coethnicity: diversity and the dilemmas of collective action, New York: Russell Sage Foundation.
  • Henrich, J., Ensminger, J., McElreath, R., Barr, A, Barrett, C., Bolyanatz, A., Cardenas J.C., Gurven, M., Gwako, E., Henrich, N., Lesorogol, C., Marlowe, F., Tracer, D., and Ziker, J. (2010) ‘Markets, Religion, Community Size, and the Evolution of Fairness and Punishment’, Science, 327: 1480–1484.
  • Herrmann, B, Thöni, C., and Gächter, S. (2008), ‘Antisocial Punishment Across Societies’, Science, 319: 1362–1367.
  • Karlan, D. S. (2005), ‘Using Experimental Economics to Measure Social Capital and Predict Financial Decisions’, American Economic Review, 95: 1688–1699.
  • Marwell, G. and Ames, R.E. (1979), ‘Experiments on the Provision of Public Goods: Resources, Interest, Group Size, and the Free-Rider Problem, American Journal of Sociology, 84: 1335–1360.
  • Ostrom, E. (2000), ‘Collective Action and the Evolution of Social Norms’, The Journal of Economic Perspectives, 14: 137–158.
  • Ostrom, E., Walker J., and Gardner, R. (1992), ‘Covenants with and without a Sword: Self-Governance is Possible’, American Political Science Review, 86(2): 404–17.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Basic HTML is allowed. Your email address will not be published.

Subscribe to this comment feed via RSS

%d bloggers like this: