Please report errors in award information by writing to: awardsearch nsf. Search Awards. Recent Awards. Presidential and Honorary Awards. About Awards.
How to Manage Your Award. Grant Policy Manual. Grant General Conditions. Cooperative Agreement Conditions. Yet past work on the leadership solution has not examined whether and to what extent leaders disproportionately affect cooperative norms in their groups via these two mechanisms, compared to peer punishment where no one individual has an enhanced role in the group.
This work demonstrates that the sanctioning institutions one experiences is critical in fostering cooperation. Generally, the institutions that govern us help shape our notions of what is right or wrong, even beyond the reach of those institutions [ 28 ]. Taken together, the aim of the present study is to further understand how leadership institutions fare, compared to peer punishment institutions, in fostering long-lasting cooperation.
Four-person groups completed a standard public good dilemma, with nine rounds of a baseline non-punishment phase and nine rounds of a punishment phase. In each round of both phases, following past work [ 8 , 9 ], participants were given an endowment of 20 monetary units MUs. They simultaneously decided how many MUs, from 0 to 20, to contribute to a group fund.
MUs contributed to the group fund were doubled and then distributed equally to all group members; MUs not contributed were kept for the self, but were not doubled. As a result, the task poses a social dilemma: there is a tension between what is best for the individual to keep all resources for the self and what is best for the group to contribute all resources to the group fund. In the punishment phase, either all group members peer punishment condition or one randomly assigned group member leader condition could punish the others.
Leaders could not be punished and had their position for the remainder of the phase. Finally, after the punishment phase, participants completed a previously unannounced one-shot version of the public good dilemma. Nor would anyone be able to administer or receive punishments. See the Methods section for more details and the S1 File for the full text of these instructions. Because the data were nested in two or three levels participants in groups for analyses on contributions in the one-shot public good dilemma, or decision rounds within participants in groups for the analyses on the repeated public good dilemma , the analyses reported here use multilevel linear regression models [ 29 ] with maximum likelihood estimation.
To ensure that fixed effect estimates were robust to violations of assumptions of multilevel models i. Round was treated as a linear variable in order to examine general contribution trends over time. But follow-up analyses treating round as a series of dummy variables given that round can range only from one to nine, it is not a continuous variable did not substantively change any findings. The S1 File contains additional details and ancillary analyses.
Fig 1A shows group-level contributions across the nine rounds of the first phase of the study: the baseline non-punishment rounds of the public good dilemma. This is the typical finding in standard public good dilemmas without punishment and demonstrates the key problem: without the presence of punishment or other incentives for cooperation, group members free-ride off of the efforts of others and contributions spiral downward dramatically. Additionally, contributions fell over time to a similar extent in both conditions i.
Taken as a whole, these results suggest that across conditions, the groups in this phase appear statistically similar. However, because the group composition remained unchanged from the non-punishment phase to the punishment phase, behavior occurring in this phase could have influenced behavior in subsequent phases. Contribution to the public good across conditions, non-punishment phase 1a and punishment phase 1b. Between the final round of the first phase and the first round of the second phase, participants were told that sanctions would now be possible.
Analyses revealed that contributions were significantly higher in the first round of the punishment phase than in the previous round, i. This was the case in both conditions: the effect of introducing sanctions did not differ across conditions i. Thus a better test of the effect of punishment is an analysis of behavior in all rounds, including those later rounds where we might observe cooperation dropping over time, as in the non-punishment phase.
The next analyses consider how the introduction of punishment affected contributions throughout the entirety of the phase. Fig 1B displays the contributions across the nine rounds of the punishment phase by condition. But unlike in the non-punishment phase where contributions declined dramatically, under the punishment institution, contributions remained high after their initial, significant increase in round one of the phase, staying consistently high in the leadership condition and increasing over time in the peer punishment condition see Table C in S1 File , Model 2.
That contributions remained high across the entirety of the punishment phase supports the notion that the introduction of sanctions promotes cooperation, as has been found in a host of past work [ 8 — 11 ]. As noted above, these results control for individual and group average contribution in the equivalent round of the non-punishment phase. Additional models containing control terms for gender, age, and pool type did not impact any of the key results; see the S1 File for these ancillary analyses.
The results on contributions appear suggestive that, overall, peer punishment and leadership were effective in promoting contributions to the public good, compared to when sanctions are not possible. But because punishments are costly, overall welfare—i. And as noted above, installing one group member with the ability to administer punishment, rather than allowing all to punish all, can also keep welfare-destroying punishments low by avoiding issues of counterpunishment and coordination. Indeed, earnings after accounting for the costs of sanctions were significantly lower in the peer punishment condition, compared to the leader condition.
On the whole, then, these preliminary analyses suggest that leadership promotes similar contribution levels, with reduced welfare destruction via sanctions, compared to peer punishment. The next section moves from the preliminary descriptive analyses of contributions and earnings to the analyses testing the arguments above: that leader behavior in particular will establish cooperative norms in groups with leaders compared to peer behavior in groups with peer punishment , and that these norms of cooperation will last in subsequent interactions.
As noted above, designating a solitary group member to administer sanctions alters the group facing the public good dilemma: it gives the leader enhanced power over the others, by allowing them the sole ability to administer sanctions. Because of the inclusion of terms for punishment received and leader contribution, leaders who could not be punished and those in the peer punishment condition who did not have a leader were omitted from these analyses, though I return to the peer punishment groups later. Highly cooperative group leaders induce cooperation in their followers in subsequent decisions; those who cooperate less are less likely to encourage cooperation among their non-leader counterparts.
Additional control terms for gender, age, and pool type did not affect results; see Table E in S1 File. Next I turn to the punishment received from the leader. Given that punishment is proposed to be more likely to promote cooperation when it is administered by high contributors to low contributors than when it is antisocial, i. To facilitate comparisons with the peer punishment condition in analyses described later given that the maximum punishment that one could receive differed in the leader condition vs.
As a result, the standardized punishment received variables could range from 0 e. Of course, this does not imply that sanctions do not promote cooperation. First, because half of the groups were omitted from these analyses, power is low. Later analyses include the peer punishment groups and allow for enhanced power to detect changes in cooperation after receiving punishment. Perhaps more importantly, the results on changes in contributions from the final round of the non-punishment phase to the first round of the punishment phase show that sanctions served as an important deterrent , promoting cooperation even in round one and throughout the remainder of the phase.
That is, the capacity to sanction, whether or not sanctions were actually deployed, promoted cooperation. But it is also clear that leader contributions are critical in shaping subsequent cooperative behavior in the groups to which they belong, beyond simply introducing the ability for a leader to punish.
After all, people generally conform to the cooperative behavior of their group members [ 36 , 37 ]. Perhaps unsurprisingly, group-level cooperative behavior in the previous round predicted own cooperative behavior in the next round. Leaders were especially salient in promoting or reducing cooperative decisions among their group members. Their choices made a larger impact than did the choices of their non-leader counterparts. It is possible that peer punishment institutions are equally effective poor at establishing cooperative norms, compared to those with leaders, if they have a high low contributor.
But another possibility is that peer punishment groups are also just as influenced by the other group members and also yield very high or very low contributions, depending on the presence of a particularly high or particularly low contributor. As a result of these issues, in follow-up analyses I matched each peer punishment group with a corresponding, unique leader group.
There was one more group in the leader condition than in the peer punishment condition, so all peer punishment groups were successfully matched with a unique leader group. This yielded a randomly chosen group member whose influence over the other members of the group could be compared to that of the influence of the leader observed in the leader groups. Table 2 displays results of models like those in Table 1 , but with the peer punishment condition included. Model 1 considers whether contributions were impacted by the behavior of a selected member of the group in the previous round—that selected member being the leader in the leader condition, or a randomly chosen peer held constant across rounds in the peer punishment condition.
But were leaders more influential than their counterparts in the peer punishment condition at establishing norms of cooperation? Model 2 demonstrates that they were. The main effects are also significant and can be interpreted as follows.
- Recommended for you.
- SUMO Protocols.
- Home | Department of Agriculture Cooperation & Farmers Welfare | Mo A&FW | GoI.
- Widening the European Union: The Politics of Institutional Change and Reform (Routledge Advances in European Politics, 5)?
- Catalog Record: Group coordination and cooperative control | HathiTrust Digital Library.
A high contributing leader is more effective than a high contributing peer at establishing high norms of cooperation in the group. Likewise, a low contributing leader is more detrimental to the group, inducing followers to contribute less. Note : Fig 2 displays estimates for contribution as predicted from Model 2, with the remaining predictors held constant at their means.
Finally, I also examined whether the effect of receiving punishment differed across conditions. Rather, in this study, the introduction of sanctions in round one and then their use throughout the punishment phase appeared to affect participants similarly, regardless of whether the punishment was administered by a peer or a leader. The next section considers contribution behavior in the one-shot public good dilemma, once participants were told that the institution was removed and punishment was no longer possible.
Note : Fig 3 displays estimates for contributions in the one-shot dilemma, as predicted from the model given in the paragraph above. After all, when making cooperative choices outside the institution, people may have simply opted to do what was working well for them in the past—that is, what they did inside the institution. These results also control for total prosocial and antisocial punishment received from the leader, which were not significantly associated with one-shot giving.
Consistent with the findings from the round-level analyses described above see Table 1 , a model regressing total leader giving on the mediator i. Importantly, as shown above, it does so above and beyond the influence of peers in peer punishment institutions. Fig 4 displays the results of the mediation analysis. The full models for these analyses are also displayed in Table F in S1 File. Mediation analyses consisting of multilevel linear regression models maximum likelihood estimation with random intercepts for the group.
Models also contained control terms for prosocial and antisocial punishments received within the institution. See also Table F in S1 File. While much past work has demonstrated that the introduction of peer punishment promotes cooperative behavior, as noted above, peer punishment faces several downsides.
Designating a single group member to administer punishment addresses several of these key downsides, but as I show here, it also gives the leader disproportionate ability to shape group norms.
You are here
This study demonstrates that, on the whole, leadership can promote cooperation to a similar extent as peer punishment. If leaders engage in cooperative behavior, their groups can achieve larger public good production than those in peer punishment institutions, while avoiding the downsides of peer punishment, including coordination problems and high costs. At the same time, leadership will fail if the leader opts not to cooperate—in these situations, peer punishment is more likely to yield successful public good production.
Of course, in this study the group member in the leader position was endowed with a special ability within the group: the ability to punish. As noted earlier, past work has suggested that granting punishment ability to one group member will fail if that group member chooses to hoard resources rather than using them to punish [ 22 ]. Taken together with the results discussed here, leaders who both cooperate and are willing to punish noncooperation—i.
That is, this work builds on work suggesting that punishment is important in promoting cooperation by suggesting that not only should group leaders be willing to use their resources to sanction group members, they must also be willing to contribute to the group, modelling cooperative choices rather than simply enforcing them in others. Because this study was conducted in anonymous four-person groups, for example, demographic characteristics were not as able to play a role as they might in real-world, face to face interactions.
Additionally, past work has demonstrated that leaders with a prosocial social value orientation are more likely to promote cooperation than those who are proself [ 19 ]. On the whole, whether the effects of the study here are moderated by these and other leader traits or characteristics could be tested via extensions of the experiment described above. Note that, like any study, this one has limitations.
- Dictionary of Medical Terms (Barrons Medical Guides).
- Case Studies in Organizational Communication: Ethical Perspectives and Practices.
- Structural aspects of building conservation.
- Kentucky Bourbon Country. The Essential Travel Guide.
- Catalog Record: Group coordination and cooperative control | HathiTrust Digital Library.
- Submission history.
One issue is that, in this experiment, the maximum amount of punishment that one could have received in the leader condition differed from that of the peer punishment condition 10 MUs in the leader condition; 10 from each of three others, thus 30 MUs total, in peer punishment. As a result, one unit of punishment may have carried greater weight in the leader condition, resulting in differing levels of punishment use and welfare destruction due to the study design.
Such a study could answer additional questions about the effects of the distribution of punishment power on cooperation that this one was not designed to address. The key difference was that the decision outside the institution was one-shot, not relayed to the group members, and free from sanctions. By relying on a highly similar behavior that participants had experienced over the course of several rounds, it is possible that part of their decision in the interaction outside of the institution was due to inertia: sticking with the same strategy they used previously.
Granted, the results stand: those with cooperative leaders behaved more cooperatively while those with cooperative peers were less affected; this enhanced cooperation predicted further cooperation even when sanctions were removed. This would yield additional insights into whether and in what contexts cooperative leadership promotes cooperative followers that are beyond the scope of the current work.
Additional considerations for future work might consider the varying ways that leadership is conceptualized in research on groups facing social dilemmas. While, as noted above, past work has examined designated punishers as authority figures or group leaders [ 16 — 21 ], others have considered different forms of leadership in these groups. It is thus far unclear whether the effect of a leader who contributes first sets an example that persists in subsequent interactions, once the leader is no longer present. Leaders who both cooperate first and are able to sanction may harness the benefits of both forms of leadership, and as in the study presented here, continue to influence their followers outside of the institution.
Sign In. Access provided by: anon Sign Out. Motion coordination through cooperative payload transport Abstract: In this paper we consider a group of agents collaboratively transporting a flexible payload. The reaction forces between the agents and the payload are modeled as the gradients of the nonlinear potentials that describe the deformation of the payload. We develop decentralized control laws without explicit communication such that the agents and the payload will eventually move with the same velocity meanwhile the contact forces are regulated.