1
Intergenerational cooperation:
motivating prosocial behavior in
favor of the distribution of finite
resources in sequential chains.
Master thesis
Social & Organization Psychology
R
oel van Vugt
in collaboration with Anne van Olffen
Master Thesis Psychology, Social & Organizational Psychology. Institute of Psychology
Faculty of Social and Behavioral Sciences – Leiden University Date:17-05-2017
Student number: 1747150
First examiner of the university: Jörg Gross
2 Abstract
Today’s behavior and deeds determine the quality of life of future generations. Worrisome topics such as climate change and exhausting energy resources such as fossil fuel raise the question whether we should take responsibility for this as humans. If the goal of society is to maintain a sustainable planet, living standard and social welfare, action must be taken to achieve this. It all depends on the level of cooperation between people to accomplish a fair distribution of finite resources and commons. One way for cooperation to emerge is through prosocial behavior and generalized reciprocity.
Generalized reciprocity is a form of prosocial behavior that explains social interactions in hierarchical structures and can be illustrated in the following way: person A helps person B, then person B helps person C. Here, we examined the role of generalized reciprocity in the distribution of a finite resource in sequential chains of people. The question is: how do we enforce cooperation through generalized reciprocity between generations? In this study we investigated whether the possibility of implementing emerging and self-imposed rules could lead to a fairer distribution of a limited resource and thus motivate prosocial behavior and cooperation. This was accomplished by mimicking the incentive structure of generations we see in the real world in a laboratory setting where participants were divided in groups of 4 and two conditions: one with and one without the possibility to implement rules. Within these groups, participants sequentially transferred units from a limited resource to their private account. We found that, contrary to the expectations, self-emerging rules did not influence the way participants distributed a limited resource. However, we did find that participants consistently kept more of the limited resource to themselves than they allowed others to do so by implementing low binding rules. This could be explained by theories about generalized negative reciprocity and psychological reactance.
3 Introduction
Predicting the future
The future and the word ‘uncertainty’ go hand in hand. Predicting the future has always been a precarious business. Though, modern day science and technology have grown fast, which, in turn, makes scientists able to say useful things about the path ahead. The path that concerns future generations: our children, grandchildren, great grandchildren, and so forth. Today’s behavior and deeds determine the quality of life of these future
generations. Different examples of well-known worrisome topics are climate change, and the underlying greenhouse effect, overpopulation and exhausting energy resources such as fossil fuel. When it comes to these topics, one might start to wonder what the planet will look like for future generations when so called intergenerational dilemmas are ignored or are dealt with in aselfish, unfair and non-cooperative way. These dilemmas are labeled intergenerational because it is not just the present generation’s quality of life that will be influenced: chances are the next generation(s) will be affected by our
behavior as well (World Commission of Environment and Development, 1987). The dilemma here is whether we will care about these future generations and thus take action, or act rather selfishly. This is the reason why intergenerational dilemmas are so important to consider. Who decides what we are obliged to leave behind for the next generation when speaking of finite resources such as fossil fuels? To distribute finite goods fairly over next generations, cooperation across generations will be needed, if the goal is to maintain a sustainable planet, living standard and social welfare (Hauser, Rand, Peysakhovich, and Nowak, 2014). According to the Brundtland Commission Report sustainable development is defined as “development that meets the needs of the present without compromising the needs of the future” (Mintzer & Michel, 2001; World
Commission of Environment and Development, 1987).
Social dilemmas
‘Public goods’ are directly related to those needs. Public goods are goods that are available to everyone. It is a resource that everyone has access to and that is non-
4 excludable1 and non-rivalrous2 (OECD, 2011). Examples of public goods are fresh air, public parks, street lightning and knowledge. Distribution and maintaining of these goods leads to a social dilemma: polluting the air does not only pollute an individual’s own air, but that of others as well. By, for example, heavily polluting the air with machines to make profit as a country, one will leave other countries and generations without fresh air. How do we deal with this? Especially fair usage of these goods is important for future generations to exist, as selfish behavior of one generation will affect the quality of life of the next generation(s) (World Commission of Environment and Development, 1987). A very famous example of a generalized abstraction of many real-world situations is known as ‘The Tragedy of the Commons’. This social dilemma is explained by Hardin (1968) and predicts the overexploitation of global public goods driven by self-interest. Hardin explains the dilemma by taking a public grassland open to all as example. This grassland (the public good) is shared by cattle of different herdsman. When one of the herdsman attempts to add more cattle to his herd, the problems start to emerge. This action will have a positive and a negative side. The positive side is that it will yield more profit to the herdsman. On the negative side, this extra animal will cause additional overgrazing of the pasture. Because this land is publicly owned, all the herdsman together will pay for the consumption done by the extra animal by leaving less grass for their own cattle. The first herdsman, however, will only reap the benefits of his action. To compensate for their loss, the other herdsman may add more cattle to the grassland themselves. This will eventually lead to overexploitation of the pasture and collapse of the ‘common’ due to overuse (Hardin, 1968). In general, social dilemmas can be explained by situations in which an individual benefits from a public good by acting selfishly (Milinski,
Sommerfeld, Krambeck, Reed & Marotzke, 2008). The group will suffer by this behavior, unless they choose to act selfishly as well. In that case, everyone will lose as the public good will be overexploited as explained by Hardin (1968). Examples of these public goods dilemmas in the modern real world are overfishing of the sea, air pollution, oil consumption and climate change. All of which are problems concerning people who
1 Non-excludable means that it is generally impossible for an individual to exclude other people from using the good.
2 Non-rivalrous means that the consumption of one individual does not prevent others from using it. Unless heavy overconsumption causes the quantity of the good to reduce in such a way it becomes unavailable.
5 handle out of self-interest. It is clear that reduction of greenhouse gas emissions to reduce the risk of dangerous climate change – about 50% of the current level by the year of 2050 – cannot be fulfilled by one country: international cooperation is required to accomplish this (Milinski et al., 2008). According to research done by Patchen (2006) in the USA about public attitudes toward climate change, most people think that more action should be taken by governments to help solve problems concerning climate change. Interestingly though, research done among 16.000 representative Europeans throughout fifteen
different countries indicated that they selected the European Union as best institution for making decisions about the protection of the environment (European Commission, 2002; Patchen, 2006). International treaties such as the Kyoto Protocol (1997) in which
currently 191 countriestake part, are an example of international cooperation to fight global warming by implementing rules to restrict behavior and reduce current emissions to protect the environment (United Nations Framework Convention on Climate Change, 2014). Another, more recent, example is the Paris Agreement which was formed in 2016 by 195 countries, with the main goal to control the increasing global temperature of the earth by fostering climate resilience and reducing greenhouse gas emissions. These types of international agreements require countries to work together and restrict their current behavior, to reach a collective goal. Rules and regulations seem to be the means to an end in these agreements because they are the point of focus within these treaties. Though, governments might want to let others do the ‘hard work’ of reducing greenhouse gas emissions (by, for example, introducing strict legislation) to benefit from the public goods themselves. So called ‘free-riders’ can be seen as the culprit of such unfairness, the core problem of social dilemmas (Olson, 1965). Where others pay the cost of collective action, riders benefit from resources, goods or services without paying. Thus, free-riders make it harder to achieve the collective target and might discourage other
individuals to participate in reaching the collective target as well (Weismuller, 2012). If society wants to find a solution for the free-rider problem, it could decide to restrict specific behavior by implementing regulations such as laws to limit the amount of fish being caught, the number of cars that are allowed on the road, congestion charges and CO2 emission per year. The downside to this is that these restrictions will have short-term negative economic effects, which makes them less likely for governments to be
6 implemented. The other side is that failing to implement these restrictions will cause dangerous climate change for later generations (Intergovernmental Panel on Climate Change, 2007; Milinski et al., 2008).
Reciprocity
The question is, how can free-riders be made to cooperate and restrict selfish behavior? There is a high volume of research about mechanisms that promote cooperation based on reciprocity, a form of prosocial behavior, coming from disciplines such as philosophy (Sangiovanni, 2007), anthropology (Fehr, Fischbacher & Gachter, 2002), sociology (Gouldner, 1960), evolutionary biology (West, Griffin & Gardner, 2007) and notably psychology (Yamagishi & Kiyonari, 2000) and economics (Bolton & Ockenfels, 2000) (Schneider, 2012, p. 36; Chiong & Kirley, 2015). Prosocial behavior is defined as any behavior that is performed by an individual, that benefits another individual and imposes a cost on the benefactor in the form of time, money or effort (Simpson & Willer, 2008). Reciprocity is a form of prosocial behavior that explains cooperation in social
interactions. Direct reciprocity occurs when person A helps person B, and person B helps person A in return, also seen as “I do something for you and you do something for me” (see Table 1) (Herne, Lappalainen, Kestilä-Kekkonen, 2013). Another form of reciprocity is called indirect reciprocity. Person C helps person A, because he or she has witnessed person A helping person B (Herne et al., 2013). Direct reciprocity is based on repeated social interactions between two individuals whereas indirect reciprocity is based on knowing the reputation of others (Nowak & Sigmund, 2013). For example, when person C sees that person A is helping person B, he or she might think of person A as a good person and thus help person A in return. Person A might have intended to perform unconditional prosocial behavior towards person B, but the helping behavior can ultimately pay-off by receiving help in return based on reputation (Herne et al., 2013). Hence, indirect reciprocity, ultimately has a selfish side to it as well. For both forms of reciprocity, identifying the other person is important for the reciprocal behavior to succeed (Saavedra, Smith, & Reed-Tsochas, 2010). However, studies have shown that being able to recognize or identifying the other individual is not always possible in real life situations (Pfeiffer, Rutte, Killingback, Taborsky, & Bonhoeffer, 2005). This would
7 mean that the previously mentioned forms of reciprocity are not always applicable when it comes to explaining cooperative behavior. However, cooperation can also be observed when reputation is not involved, nor the identification of another individual from
previous interactions. For example, when a stranger gives money to a homeless individual and this homeless person in turn copies this behavior by helping an older individual cross the street. Researchers introduced a way of explaining cooperation in these situations called generalized (indirect) reciprocity (Chiong & Kirley, 2015; Pfeiffer et al., 2005). This approach is different to direct and indirect reciprocity, because it does not rely on the ability of humans to identify other individuals. Generalized reciprocity can be explained as follows: person A helps person B, then person B helps person C.
Generalized indirect reciprocity, however, is when person B helps person C, while person A has witnessed this helping behavior and thus helps person D (see Table 1) (Chiong & Kirley, 2015). It is important to consider that this behavior is based on mimicking (observed) behavior and follows a hierarchical interaction structure. These two forms of reciprocity depend entirely on observations (Rutte & Pfeiffer, 2009), or prior experiences (Pfeiffer et al., 2005).
Table 1 – Various types of reciprocal behavior (based on Chiong & Kirley, 2015).
Personal information Socially acquired information
Individual-specific information Direct reciprocity A à B B - à A Indirect reciprocity C: A à B C - à A Individual-unspecific information Generalized reciprocity B à A A - à C
Generalized indirect reciprocity A: B à C A - à D Note: For direct and indirect reciprocity, individual-specific information is required. In contrast, forms of generalized reciprocation use individual-unspecific information. With direct reciprocity, person B helps person A because person A has helped B in the past, repeated encounters between individuals A and B are required for this. As for indirect reciprocity, person C helps person A because person C has seen person A (good reputation) helping person B. In the case of generalized reciprocity, person B who has helped person A, doesn’t necessarily receive help back. Person A will help person C because he/she is grateful for being helped out previously by person B. As for generalized indirect reciprocity, the interactions between the individuals can be completely independent. Person A could help person D, solely because person A has witnessed the cooperation between person B and person C.
There are plenty of examples of generalized (indirect) reciprocity in daily life. On the internet, for example, people upload their code for other people to use, adjust it, and pass it on in open source projects. Other examples are online discussion boards or websites containing free information such as various encyclopedic data bases (see Chiong &
8 Kirley 2015 for an overview). The distribution of certain services can also be seen as a form of prosocial behavior. There are different motivations to perform prosocial behavior among which is altruism. On the definition of altruism (and even the existence of it) is very little agreement across scientific disciplines (Simpson & Willer, 2008). Though, one definition that is used along different research is as follows: behavior to increase another individual’s welfare without expecting anything in return, even at one’s own costs. (Simpson & Willer, 2008; West, Griffin, & Gardner, 2007). This is where altruistic behavior differs from direct and indirect reciprocity, because these forms of behavior are either network or circle based where altruistic behavior is not (Boyd & Richerson, 1989; Greiner & Levati, 2005). In network or circle based interactions, behavior ends where it started. With (in)direct reciprocity, for example, the initiator of helping behavior will eventually receive help in return. They are based on mutual acts of cooperation, whereas acts of altruism are not initiated to benefit from in any way (Simpson & Willer, 2008; Greiner & Levati, 2005; West et al., 2007). Generalized (indirect) reciprocity, however, comes closer to altruism as this type of reciprocal behavior does not require a two-way interaction, it is based on a hierarchical interaction structure.
Considering the social dilemmas discussed in the beginning of this paper (public goods dilemmas such as climate change and fossil fuels), one can debate on which type of behavior is applicable in the search for ways to motivate cooperation. Different forms of reciprocity seem to be beneficial when it comes to encouraging cooperative behavior as described earlier. But when it comes to intergenerational cooperation direct and indirect reciprocity cannot explain cooperative behavior, because they assume the
existence of mutual cooperation and identification of previous interaction partners which is not possible between generations as they do not exist anymore. For instance,
generation A can be nice to generation B (by, for example, restricting their lifestyle in a certain way) and B can be nice to generation C, but generation A will never directly or indirectly profit from their own performed cooperative behavior by generation B or C. Which is one of the aspects of direct and indirect reciprocity: the (indirect) aim to benefit from the performed behavior (Simpson & Willer, 2008). Therefore, it is tempting for generations to free-ride, as there will be no direct consequences. In this hierarchy structure of intergenerational interactions, therefore, a form of generalized (indirect)
9 reciprocity or altruism combined with rules or binding contracts might be required for cooperation to emerge, because these types of motivations to perform prosocial behavior do not rely on two-way beneficial interactions.
Present study
Nowadays, one of the most determining factors for the welfare of future generations is the way in which global public goods, such as finite resources, are treated and the extent to which they are distributed over future generations. As mentioned, two forms of reciprocity could play a part in this. The mentioned types of interactions, generalized (in)direct reciprocity), are similar to the way generations interact with each other: a hierarchical structure of one-way interactions. Generation X will never be able to receive anything in return from generation X+1. It is imaginable to conceptualize the interactions between generations as a sequential chain of interactions, though the question is: how does cooperation emerge within such chains? We are not aware of any past research that has investigated certain behavior. Therefore, in this study, we will examine whether and how cooperation emerges in the distribution of a finite resource in sequential chains of people. To investigate how cooperation can be sustained in these chains is crucial to maintain a sustainable planet. Without cooperation between generations, quality of life of future generations will be in danger (Heath, 2013). Observing the way individuals behave in sequential chains might help us to understand how groups and societies behave in similar situations. Our hypothesis is that the finite resource will be distributed unequally in such a way that the participants will act rather selfishly for reasons explained by ‘The Tragedy of the Commons’. People will tend to act out of self-interest instead of
contributing to the collective interest, because there is no possibility of (in)direct reciprocation and thus no potential future benefit. If generation A is treated unfairly by generation B, A can not negatively reciprocate next time. Thus, the question arises how to combat such unfairness in these sequential chains? Rules and regulations can be a way to restrict behavior. However, these rules should be self-emerging and self-imposed similarly to the Kyoto Protocol and the Paris agreement, as general rules about the distribution of finite resources do not exist. Though, acting in a self-interested way will always be more beneficial to the individual. Which in turn raises the question whether
10 these rules will even emerge? Therefore, we will investigate if the possibility of
implementing self-emerging and self-imposed rules can lead to a fairer distribution of a limited resource. We expect that the possibility of being able to impose rules will create a more equal distribution of the finite resource, and that the possibility of making self-imposed rules encourages prosocial behavior and cooperation. In addition, we expect that the level of fairness depends on the amount of rules that have been implemented. In this experiment we try to mimic the incentive structure of generations we see in the real world, by investigating it more closely in a laboratory setting.
Method
Recruitment
In the period between February and May 2017, 184 students between the age of 17 and 58 (M = 21,76, SD = 3.83) participated in this experiment. Different recruitment
strategies have been used to find participants for the experiment. An advertisement was placed on the online recruitment platform of the University website in order to find students in need of course credits. Also, flyers were distributed among students at Leiden University. The experiment was advertised with information concerning time and
rewards. The experiment lasted for a maximum of 15 minutes and participants could earn 1 credit or €3,-. In addition, they could earn up to €10,-, depending on their decisions in the experiment.
Measurement procedure
The experiment was conducted in a lab setting at Leiden University. The material used for this experiment was a computer program containing the complete experiment,
including instructions, comprehension questions and a demographic questionnaire. At the beginning of the experiment, participants were assigned to a group of four. Half of the participants were randomly allocated to a group in the no-rule condition and the other half to a group in the rule condition. In both conditions, participants started with reading the instructions on the computer screen (see Appendix A.1), detailing the rules and procedure of the experiment. To ensure proper understanding, participants had to answer
11 comprehension questions (see Appendix A.2). The participants could only proceed to the experiment when all answers were filled in correctly.
Design
In both conditions, the four participants were asked to sequentially transfer units from a pool of 100 units to their private account, one after another. Hence, participant 1 was asked to transfer units to his or her private account, then participant 2 was asked to transfer units to his or her private account from what was left in the pool, then participant 3 and so on. Each unit had a value of 0.10 eurocents. The units that the participants transferred to their private account were exchanged to euros and paid out accordingly after the experiment was over. Before transferring units to their private account, each participant was informed about how many units were left in the pool, and how many participants would follow after them in the group. In the rule condition, however, there was one difference: after the participants filled in the amount of units he or she wanted to transfer to the private account, the participants had the option to impose an upper limit on how much the next person could transfer from the pool to his or her private account. For example, participant B saw that there were 70 units left in the pool and then decided to transfer 30 units to his or her own account and set a binding rule for participant C limiting his or her maximum transfer amount to 20. Participant C now could transfer a maximum amount of 20 units to his or her private account and set a rule for the next participant, if he or she wanted to. The independent variable in this experiment was thus the option to impose an upper limit on how much the next person could transfer from the pool to his or her private account which was present in the rule condition, but absent in the no rule condition. With this variable, the option to impose a binding rule, we tried to manipulate the amount of units (dependent variable) that was transferred to the private accounts of the participants.
Data analysis
The data was analyzed with SPSS (Statistical Package for the Social Sciences, version 23). For each group (e.g. group 1, participants 1 to 4) of the experimental (rule condition) and control condition (no rule condition), the transfer amount means, the standard
12 deviation of transfer amounts per group per condition, the relative transfer means and equality deviation means were calculated. The standard deviation of transfer amount means explains the difference of transfer amount between the group mean and overall mean. The relative transfer mean describes the percentage of units participants transferred to their private account from the remaining units per mover (e.g. 20 out of 60 units would be 33%). The equality deviation is a measurement that tells how much the transfer
amount deviates from a fair transfer amount (25 units per person). Mann-Whitney U tests were performed to measure differences in standard deviation across conditions per group, to measure differences in relative transfer means between group members across
conditions, to measure differences between equality deviation across conditions and to measure differences between equality deviation in the experimental condition between rule and no-rule implementers. A linear regression was performed to predict equality deviation from the amount of rules that were implemented and a paired sample t-test was performed to compare transfer amount means with rule amount means.
Results
To test our first hypothesis, we looked at the mean transfer amount of
participants per group, per condition (see Figure 1). As can be seen, the mean transfer amount for the first group member is nearly identical across conditions (no rule condition M = 40.48, SD = 22.03 and rule condition M = 40.35, SD = 22.90). The mean transfer amount between the second (no rule condition M = 25.22, SD = 18.72, rule condition M = 20.61, SD = 12.64) and fourth (no rule condition M = 13.30, SD = 12.17, rule condition M = 18.09, SD = 14.07) movers across conditions is different where the mean transfer amount of group member 3 (no rule condition M = 19.22, SD = 13.36, rule condition M = 18.70, SD = 9.93) is nearly identical across conditions. The reference line in Figure 1 shows what the fair allocation benchmark: 100 units divided by 4 people (i.e. 25 units per group member). As can be seen, the actual allocation deviates from the fair allocation benchmark across both conditions. Only group member 2’s deviation in the no rule condition shows a nearly fair allocation.
13
Figure 1 - Mean transfer amount per group member in rule and no rule condition. Reference line fair allocation
benchmark: 25.
The mean transfer amount of all groups together, however, did not give us information about the differences of transfer amounts within groups. This information was needed to be able to compare separate groups with each other, per condition. Therefore, the
standard deviation of transfer amount per group per condition was calculated. The fourth mover data was left out during this comparison, as this participant would generally take all the units that are left in the pool and is not able to implement binding rules. This makes their data invaluable for what we try to investigate in this experiment. As can be seen in Figure 2, the standard deviation (the difference of transfer amount between group mean and the overall mean) is slightly higher for the rule condition (M = 17.93)
compared to the no-rule condition (M = 14.69), indicating that groups in the no rule condition distribute units fairer than groups in the rule condition. However, the difference in fairness deviation was not significant across conditions (Mann-Whitney U Test U = 210.500, p = .234).
14
Figure 2 - Mean standard deviation of transfer amount, as a measure for inequality in transfer amounts, per group per
condition (N = 23 per condition) (error bars show the 95% confidence interval).
Because the comparison of standard deviation of transfer amount per group between conditions did not capture independent transfer amount per participant within groups, the relative transfer amount was calculated in order to find out more about the differences in fairness of transfer amount per group and per condition. The relative transfer amount can be explained as follows: an equal split of the units means that the first mover would take 25% of the total amount of units, the second mover would take 33% of the amount that is left in the pool and the third mover would take 50% of what’s left in the pool, leaving the other 50% for the fourth mover. To illustrate: if the first mover decided to transfer 60 units to his/her private account, leaving 40 units for the next mover, his/her relative transfer amount would be 60%. If the second mover in this group decided to transfer 30 units to his/her private account, his/her relative transfer amount would be 75%. The relative transfer amount thus is a measure of fairness in transfer amounts independent of the previous movers. Again, the data of group member 4 was left out in this comparison, because information about the relative transfer amount of the last group member is not valuable to our study results. The mean relative transfer amount per group member per condition can be seen in Figure 3. The graph shows that the mean of group member 2 (M = 33.54%) and group member 3 (M = 49.29%) in the rule condition is nearly similar to an equal split: 33% and 50%. The participants in the no rule condition generally transfer more to their private account than the equal split amount (group member 2 M = 44.95%, group member 3 M = 55.97%). However, looking at the graph, the first movers in both
15 conditions had similar relative transfer amounts (no rule condition M = 40.48%, rule condition M = 40.35%). Though, none of these differences were statistically significant (Mann-Whitney U test: group member 1 U = 257, p =.864, group member 2 U = 201.500, p = .459, group member 3 U = 167.500, p = .253).
Figure 3 - Mean relative transfer amount per group member per condition. Reference lines at equal split 25%, 33% and
50% (error bars show the 95% confidence interval).
Because tests on differences in relative transfer amounts within groups did not capture differences of fairness per condition, only per group member, the equality deviation was calculated to find out more about the inequality of transfer amounts across conditions. An equality deviation of 0 represents an equal distribution of the units. For example, when group member 1 takes 25 and leaves 75 for the rest, or when group member 2 finds 60 units left in the pool and takes 20, leaving 40 for the next 2 group members. If, however, group member 1 transfers 65 units to his private account, he or she has an equality deviation of 40 because 65 units are 40 more units than a fair amount: 25. The difference between equality deviation and relative transfer amount is that the outcome of the
equality deviation is standardized for every group member. This makes it possible to compare conditions with each other, instead of group members per conditions. Again, data of group member 4 was left out in this comparison, because information about the equality deviation of the last group member is not valuable to our study results. In Figure 4 it can be seen that the mean equality deviation of the rule condition is slightly lower (M = 3.25) than the mean equality deviation of the no rule condition (M = 5.28). However,
16 differences between the equality deviation of the no-rule condition and the rule condition were not significant (Mann-Whitney U Test U = 2126.500, p = .259).
Figure 4 - Mean equality deviation per condition (error bars show the 95% confidence interval).
As previous tests focused on both conditions, no information about rule implementation and fairness of unit allocation was captured. To get a closer look at the distribution of units within the experimental condition, the role of rule-implementation in fairness of distribution of units was investigated. Did rule implementation influence fairness in transfer amounts? In other words: did participants who implemented a rule generally transfer units in a more or less fair way than the participants who did not implement a rule within the same condition? 42.02% out of 69 participants (29) who had the option to implement a rule did actually implement a binding rule. Again, data of group member 4 was left out in this comparison, because information about the rule implementation of the last group member does not exist and thus is not valuable to our study results. It can be seen in Figure 5 that participants who implemented a rule have a slightly lower equality deviation (M = 4.45) than participants who did not (M = 5.54). As a comparison, the average equality deviation (M = 7.64) for the no rule condition was added in the graph. This means that participants who implemented a rule generally transfer units to their private account in a fairer way than participants who did not implement a rule. However, a statistical test with equality deviation as dependent variable and rule implementation as independent variable indicated that the difference between rule implementation was not significant (Mann-Whitney U-test U = 546.500, p = .664).
17
Figure 5 - Mean equality deviation of participants who did or did not implement a rule. Reference line at the
mean (M = 7.64) of the control condition (error bars show the 95% confidence interval).
As the previous test only focused on whether or not a rule was implemented and the equality deviation herein, we wanted to find out whether participants who implemented a rule transferred equal amounts of units to their private account (transfer amount) in contrast to the binding rule they imposed for the next participant (rule amount). The average transfer amount (M = 27.41, SD = 13.94) is higher than the average rule amount (M = 19.93, SD = 6.30). Participants who implemented a rule generally transferred more units to their private account than the limit they imposed on the next mover (Paired Samples T-test t(28) = 2.587, p < .05).
In addition, to find out whether the amount of rules implemented per group had influence on the fairness of the distribution of units within each group, differences between the amount of rules that were implemented per group and the equality deviation were analyzed (see Figure 6). As can be seen, groups that implemented no rule had a higher average equality deviation (M = 9.53) than groups who had implemented 1 (M = 3.54) or 2 (M = 3.49) rules. None of the groups implemented 3 rules. The control group (no rule condition) was added in the graph for comparison. It shows that participants in the no rule condition (M = 10.58) have similar average equality deviation as groups that
18
Figure 6 - Mean equality deviation of groups who did or did not implement a rule (error bars show the 95% confidence
interval).
To predict equality deviation from the amount of rules that were implemented, a linear regression was conducted with number of rules implemented as the predictor. The result was, however, not significant (t(1, 21) = -1.206, p = .241 with an R² = .065).
Discussion
The current research provides new insights into the influence of self-emerging and self-imposed rules on the motivation of prosocial behavior and generalized reciprocity. Using a limited resource in the form of monetary units to be distributed in sequential chains of people, we showed that the possibility of implementing emerging and self-imposed rules did not influence fairness levels on the allocation of units. Neither in the control, nor in the experimental group could people achieve a fair allocation of the units.
This was investigated by conducting an experiment to analyze the level of fairness in the distribution of these units. The experiment was carried out in an anonymous experimental setting where participants in the control group (no rule condition) had the option to transfer units from a pool of 100 units to their private account. Participants in the experimental condition went through the same procedure though with one difference: they had the option to implement a binding rule for the next mover in the chain. This binding rule consisted of a maximum amount of units this next participant would be able
19 to transfer to his or her own private account. It was expected that participants in the no rule condition would act rather selfishly compared to the participants in the rule condition by transferring more money to their private account than what would be a fair amount: leaving 25 units for every participant. However, fairness deviations across conditions in multiple measurements indicated that there were no significant differences between the two conditions. As we expected that participants in the rule condition would allocate the units more fairly in comparison to the no rule condition, it is, hence, clear that our first hypothesis was not supported by the data. Participants did not differ in the amount of units they distributed to their private account across conditions. Another expectation was that a fair distribution of the units in the experimental condition would depend on the amount of rules being implemented by a group. This hypothesis, however, was also not supported by the data. We found no significant difference between the amount of rules being implemented by the groups and fairness levels of unit allocation. A possible explanation of this might lie in the rule condition. Namely, within the differences between the amount of units rule implementers transferred to their private account and the limit they set for the next mover. Statistical tests proved that participants consistently transferred more units to their private account than they allowed others to transfer to their private account by implementing a lower binding rule for the next participant. This behavior could be explained by theories about generalized negative reciprocity. Gray, Ward and Norton (2014) found that the concept of ‘paying it forward’ works in both ways: positively (generalized reciprocity) and negatively. When it was not possible for participants to reciprocate greedy behavior performed by previous participants towards the sender, participants might have reciprocated towards the next mover. The only way to accomplish this was to restrict the maximum amount of units the next mover could transfer from the pool to their private account (Wade-Benzoni, Hernandez, Medvec & Messick,2008). Considering the decreasing mean transfer amounts per subsequent mover (see Figure 1), there is a high chance people found themselves treated unfairly and thus reciprocated negatively to the next mover by imposing a low binding rule. Another explanation for the fact that no significant results were found concerning differences in fairness between conditions could be that people might be rule reluctant, not willing to implement rules to restrict others. In the experiment, 59.8% of the people who could
20 implement a rule did not do so. The psychological reactance theory (Berm, 1989;
Fogarty, 1997) can be referred to when trying to explain certain behavior. It describes the way individuals are motivated to reclaim their affected freedom and preventing the loss of freedom for others, when their own perceived freedom is threatened. In this case, it could explain why people did not implement rules that restrict the behavior of the next mover in the chain. This self-imposed binding rule would restrict the freedom of participants and thus might have caused participants to reclaim this freedom by not implementing rules. It also explains why people would not reciprocate the behavior of rule implementation, as opposed to what was expected, for the same reasons. By finding themselves limited in their freedom, they might have not been willing to set a binding rule for the next mover. It could, however, be an interesting topic for future research to investigate the relationship between psychological reactance and generalized reciprocity. Further, the lack of support for our hypothesis could be explained by the type of
motivation participants had to participate in the experiment. According to the demographic questions that participants had to fill out, 81% of the participants
participated in the experiment to earn money instead of participating for ‘science’. The experiment was advertised with information about time (15 minutes) and rewards
(money) which might have attracted students that participated for the sole fact of earning money quickly, acting selfishly no matter the condition they were in.
Concludingly it can be assumed that self-emerging and self-imposed binding rules do not influence levels of fairness in the distribution of a finite resource in sequential chains of people, at least not in the way it was investigated in this study. It would be interesting to conduct more research regarding this matter, as information regarding motivations for people to perform prosocial behavior in hierarchy-like interaction structures could be related to intergenerational interactions. This information, in turn, could be used to find a solution for the intergenerational dilemmas we have to deal with nowadays, such as the fight against climate change and the depletion of fossil fuels. It is clear that more research should be conducted to find a way to motivate intergenerational cooperation to maintain a sustainable living standard, welfare and planet.
21 References
Bolton, G. E., Ockenfels, A. (2000). ERC: A theory of equity reciprocity, and competition. American Economic Review, 90(1), 166-193.
Boyd, R., & Richerson, P. J. (1989). The evolution of indirect reciprocity. Social Networks, 11, 213-236.
Bicchieri, C., & Xiao, E. (2009). Do the right thing: but only if others do so. Journal
of Behavioral Decision Making, 22(2), 191-208.
Brehm, J. W. (1989). Psychological reactance: Theory and applications. NA-Advances in Consumer Research, 16.
Chiong, R., & Kirley, M. (2015). Effects of iterated interactions in multi-player spatial evolutionary games. IEEE Transactions on Evolutionary Computation, 16(4), 537- 555.
European Commission (2002). The environment? What Europeans think. Retrieved from: bookshop.europa.eu
Fehr, E., Fischbacher, U., Gachter, S. (2002). Strong reciprocity, human cooperation, and the enforcement of social norms. Human Nature, 13(1), 1-25.
Fogarty, J. S. (1997). Reactance theory and patient noncompliance. Social Science & Medicine, 45(8), 1277-1288.
Gouldner, A. W. (1960). The norm of reciprocity: a preliminary statement. American Sociological Review, 25(2), 161-178.
Gray, K., Ward, A. F., & Norton, M. I. (2014). Paying it forward: Generalized reciprocity and the limits of generosity. Journal of experimental psychology: General, 143(1),
22 247-254.
Greiner, B., Levati, M. V. (2005). Indirect reciprocity in cyclical networks. Journal of Economic Psychology, 26, 711-731.
Hardin, G. (1968). The tragedy of the commons. Science, 162, 1243-1248.
Hauser, O. P., Rand, D. G., Peysakhovich, A., Nowak, M. A. (2014). Cooperating with the future. Nature, 511, 220-223.
Heath, J. (2013). The structure of intergenerational cooperation. Philosophy & Public Affairs, 41, 31-66.
Herne, K., Lappalainen, O., & Kestilä-Kekkonen, E. (2013). Experimental
comparison of direct, general, and indirect reciprocity. The Journal of Socio-Economics, 45, 38-46.
Intergovernmental Panel on Climate Change (2007). Climate Change 2007: Impacts, Adaptation and Vulnerability. Summary for Policymakers. Contribution of Working Group II to the Fourth Assessment. Retrieved from:
https://www.ipcc.ch/pdf/assessment-report/ar4/wg2/ar4_wg2_full_report.pdf
Milinski, M., Sommerfeld, R. D., Krambeck, H. J., Reed, F. A., & Marotzke, J. (2008). The collective-risk social dilemma and the prevention of simulated dangerous climate change. Proceedings of the National Academy of Sciences of the United States of America, 105(7), 2291-2294.
Mintzer, I., & Michel, D. (2001). Climate change, rights of future generations and intergenerational equity: an in-expert exploration of a dark and cloudy path. International. Journal Global Environmental Issues, 1(2), 203–222.
23 Nowak, M. A., Sigmund, K. (2005). Evolution of indirect reciprocity. Nature, 437, 1291-
1298.
OECD. (2011). Evaluating policies for Delivering Agri-environmental Public Goods. Retrieved from: http://www.oecd.org/tad/sustainable
griculture/48185525.pdf
Olson, M. (1965). The logic of collective action; public goods and the theory of groups. Cambridge, Mass: Harvard University Press.
Patchen, M. (2006). Public attitudes and behavior about climate change. PCCRC Outreach Publication. Retrieved from: https://www.purdue.edu/discoverypark/
climate/resources/docs/Patchen%20OP0601.pdf
Pfeiffer T., Rutte C., Killingback T., Taborsky M., & Bonhoeffer S. (2005). Evolution of’ cooperation by generalized reciprocity. Proceedings of the Royal Society of
London: Biological Sciences, 272, 1115–1120.
Rutte C., & Pfeiffer T. (2009). Evolution of reciprocal altruism by copying observed behaviour. Current Science, 97(11), 1573–1578.
Saavedra, S., Smith, D., & Reed-Tsochas, F. (2010). Cooperation under Indirect Reciprocity and Imitative Trust. PLoS ONE, 5(10), e13475.
Sangiovanni, A. (2007). Global Justice, Reciprocity, and the State. Philosophy & Public Affairs, 35(1), 3-39.
Schneider, K. (2012). Becoming oneself, dimensions of 'Bildung' and the facilitation of personality development. Berlin: Springer VS.
24 Simpson, B., Willer, R. (2008). Altruism and Indirect Reciprocity: The Interaction of
Person and Situation in Prosocial Behavior. Social Psychology Quarterly, 71, 37- 52.
United Nations Framework Convention of Climate Change: Kyoto protocol. (2013). Retrieved from: http://unfccc.int/kyoto_protocol/items/2830.php
Wade-Benzoni, K. A., Hernandez, M., Medvec, V., & Messick, D. (2008). In fairness to future generations: The role of egocentrism, uncertainty, power, and stewardship in judgments of intergenerational allocations. Journal of Experimental Social
Psychology, 44(2), 233-245.
Weismuller, J. P. (2012). Social movements and free riders: Examining resource
mobilization theory through the Bolivian water war. The Macalester Review, 2(2), 4.
West, S. A., Griffin, A. S., & Gardner, A. (2007). Social semantics: altruism, cooperation, mutualism, strong reciprocity and group selection. Journal of Evolutionary Biology, 20, 415–432.
World Commission on Environment and Development. (1987). Our Common Future. Retrieved from:
http://www.un-documents.net/our-common-future.pdf
Yamagishi, T., & Kiyonari, T. (2000). The group as the container of generalized reciprocity. Social Psychology Quarterly, 63(2), 116-132.
25 Appendix A.1
INSTRUCTIONS PAGE 1
Dear participant,
In this part of the experiment, you are paired with three other participants who are also taking part in this study.
Each participant will make a decision that will determine their payoff for this part of the study, as well as influence the payoff of the other participants.
The decisions are made sequentially.
Each of you is assigned an individual number from 1 to 4 by chance.
This number will determine which participant will make the first, the second, the third, and the fourth decision.
For example, if you are number 3, two participants before you already made a decision and after you make your decision, there will be one more participant, participant number 4, who will make the decision after yours.
What the decision entails is explained on the next page.
PAGE 2
In this part of the experiment, there is a limited pool of units.
This pool will be distributed among all four participants you are paired with, according to the individual decisions of each of you.
As explained, participant number 1 will go first .
He/she will learn about how many units are in the pool, and then decide how many units to take for him/herself.
26 Whatever participant number 1 decides to leave in the pool will be passed on to participant number 2.
Participant number 2 will then learn about how many units are left in the pool, and decide how many units to take for him/herself.
Whatever participant number 2 decides to leave in the pool will be passed on to participant number 3.
Participant number 3 will then learn about how many units are left in the pool, and decide how many units to take for him/herself.
Whatever participant number 3 decides to leave in the pool will be passed on to participant number 4.
Participant number 4 will then learn about how many units are left in the pool, and decide how many units to take for him/herself.
The experiment ends with the decision of participant number 4.
PAGE 3
At the start of the experiment, you therefore will learn: – in which position you are (from first to fourth) – how many units are in the pool
– how many participants will follow after your decision
You will then decide how many units to take for yourself. You will also see how many units will be left in the pool.
COMPREHENSION QUESTIONS
For this example, assume there are 60 units left in the pool. Participant number 3 decides to take 50 units. How many units are left in the pool for participant number 4?
- 50 - 10 - 60
27 For this example, assume there are 40 units left in the pool. Participant number 2 decides to take 12 units. Participant number 3 decides to take 10. How many units are left in the pool for participant number 4?
- 18 - 12 - 50
If you are participant number 2, how many participants before you already made a decision?
- 0 - 1 - 2 - 3
If you are participant number 2, how many participants after you still have to make a decision?
- 0 - 1 - 2 - 3
28 Appendix A.2
INSTRUCTIONS
PAGE 4 (EXPERIMENTAL)
While deciding how much units to take from the pool, each participant, except for participant number 4, will make one additional decision.
This decision is about implementing a binding rule for the next participant in the form of: “The next person shall maximally take up to X units”.
Thus, participant number 1 can decide how much participant number 2 shall maximally take from the pool. Participant number 2 can do the same for number 3, and number 3 for number 4.
If a participant decides to implement the rule, he/she will also decide on X. The next participant is then is restricted to maximally take X from the pool.
PAGE 5
The amount of units you have at the end of the experiment will be transformed into money and paid to you accordingly (the same is true for the other participants). Each unit will be worth 10 euro cents.
This is the end of the instructions for this part of the experiment.
Press the button below if you understand the instructions and want to proceed.
COMPROHENSION QUESTIONS For experimental condition:
For this example, assume there are 50 units left in the pool after Participant number 2 decided how much to take for him/herself. Then, participant number 2 decides to implement a rule in the form: “The next person shall maximally take up to 20 units”. How many units can participant number 3 maximally take?
29 - 0
- 10 - 20 - 50
For this example, assume there are 50 units left in the pool after Participant number 2 decided how much to take for him/herself. Then, participant number 2 decides to implement a rule in the form: “The next person shall maximally take up to 30 units”. How many units can participant number 3 minimally take?
- 0 - 10 - 20 - 50
