Social preferences in the lab, in the field, and online


Ressource 1Download: Thesis DK-OK.pdf (12685.87 [Ko])
State: Public
Version: After imprimatur
Serval ID
PhD thesis: a PhD thesis.
Social preferences in the lab, in the field, and online
Kistler Deborah
Thöni Christian
Institution details
Université de Lausanne, Faculté de droit, des sciences criminelles et d'administration publique
Faculté de droit et des sciences criminelles
Université de Lausanne
CH-1015 Lausanne

Publication state
Issued date
Nowadays many scholars agree that the self-interest hypothesis, which postulates that humans care only about their own utility, is not the most accurate assumption to explain decision-making. Countless empirical studies provide evidence that humans have social preferences, meaning that they care about the well-being of other, unrelated individuals. Furthermore, in contrast to the predictions of standard economic theory, concerns about fairness and reciprocity affect behavior in many situations (Fehr & Schmidt, 2003). These social preferences are the central theme of my thesis.
For the functioning of modern societies, social preferences are crucial (see for example Ostrom, 1990; Bowles & Gintis, 2011). They facilitate for instance cooperation in groups of strangers, the provision of public goods or the maintenance of social order, to name just a few. Generally speaking, in many situations, the welfare of the group depends on the individual willingness to behave prosocial. In these situations, individuals have an incentive to free-ride on the others to increase their private benefit. However, if all individuals would act like this, everyone would be worse off. Let’s consider the classical example of a group work at the university. If one student free rides and the others behave prosocial and cooperate, the free rider has the highest payoff. He does something else, which increases his utility while the others work on the group project. In the end, however, he still benefits from a good grade thanks to the effort of the others. In contrast, if all group members would free ride the group work would not be done, and hence everyone would be worse off. Charitable giving or volunteering activities are other examples where prosocial behavior is necessary. Thus, for social scientists, it is important to understand the development and determinants of social preferences.
This thesis contains three essays, in which I elaborate on social preferences from dis- tinct angles. Common, besides the topic, is the methodology of experiments, which I applied throughout the thesis. The different types of experiments which we or I con- ducted rely all on standard methodologies from experimental economics. As the title of the thesis suggests, we conducted one experiment online, one in the laboratory and one in the field. Experiments can be used to either measure preferences or to elaborate factors which causally affect preferences. In chapter one, the main purpose of the experiment was to measure social preferences and their relationship to other socially relevant variables. The results of such correlational studies can reveal patterns which serve as hypothesis for subsequent research. For example, to test whether the observed relationship is causal. Furthermore, based on these correlations one can make predictions. Lastly, such correla- tions can be evaluated across countries and cultures, to elaborate cross-cultural differences. In contrast, in chapter two and three the aim was to evaluate factors which affect social preferences causally. These designs allow to draw conclusions about the determinants of social preferences. Moreover, insights from such analyses can be very valuable for policy making.
To be more precise, in chapter one, we study the link between moral values and vari- ous forms of prosocial behavior. Since decades, cross-cultural psychology examines moral values using data from standardized surveys, assuming that values guide human behavior. However, so far, the claim that moral values influence prosocial behavior has only been demonstrated for activities that respondents self-reported in surveys (Welzel & Deutsch, 2012) but never for directly observed behavior. Moreover, we are interested in a particular set of moral values; namely, emancipative and secular values. Emancipative values reflect the appreciation of equal freedoms and secular values the depreciation of sacred author- ity. They become more important as the living conditions of individuals improve (Welzel, 2013). To fill this gap, we conducted online behavioral experiments with participants from the sixth wave of the World Values Survey (WVS) in Germany. This allows us to link the respondents’ moral values, as measured by the WVS, to the same respondents’ prosocial behavior, as observed in the behavioral experiment. In the online behavioral experiment, participants conducted several incentivized decision tasks. Concretely, we elicited measures for cooperation in a public goods game, the ability to coordinate in a property rights game, and altruism using a donation decision. We test the hypotheses that, (a) emancipative val- ues show a strongly positive association with prosocial behaviors, (b) secular values show a modestly positive association, (c) at the same as they associate strongly negatively with protectionist behavior. The evidence boils down to three findings. Emancipative values relate to higher common pool contributions and larger donations to charitable organiza- tions. Secular values, on the other hand, are linked to more productive and less protective investments. As these results conform to key theories and reach empirical significance in a major postindustrial nation, we conclude that we have significant evidence at hand high- lighting the potential of combined survey-experiment methods to establish value-behavior links that are otherwise inexplorable.
In the second chapter, we conducted a laboratory experiment to evaluate the effect of salience on the contributions in a public goods game. According to standard theory in economics, an individual should process all available information before making a decision. This does not mean that all choice attributes receive the same weight in the decision-making process, but they should at least be considered to determine their weight. However, for instance, when I go to Montreal, and I want to buy something, I constantly valuate the items to be cheaper than they actually are. I neglect the taxes because the price tag does not include them. In such cases, the tax is not salient, and salience guides my choice. If and how salience affects choices is the topic of the second chapter. In the first part of the chapter, we introduce a salience factor to the Fehr-Schmidt model (Fehr & Schmidt, 1999) to derive theoretical predictions. All bilateral comparisons that individuals make are weighted by how salient the behavior of other relevant group members is. Based on this extension, we hypothesize that the salience of the contributions of the other players should affect contributions in a standard public goods game systematically. We implement two treatments to test our predictions empirically. In the maximum treatment, the highest contribution is most salient. In contrast, in the minimum treatment, the lowest contribu- tion within a group is most salient. Our results are surprising: We find that people do not adjust their contributions according to our hypotheses. If the lowest contribution in a group is most salient, it does not lead to an immediate decline in the contribution. Vice versa, if the maximum contribution is salient, it does not result in higher contributions. Based on our results we hypothesize that focusing on the maximum provides an upper bound of the acceptable contribution level, while the minimum serves as lower bound.
Lastly, in chapter three, I elaborate how the social environment of children affects their development in general and in particular the formation of their preferences. Children from low-income families often perform worse with respect to skills compared to children from more advantaged families. Research on early childhood development suggests that providing children with high-quality preschool education potentially closes this skill gap (Heckman, 2006). I elaborate on an intervention which aimed at improving the quality of care in community nurseries in Colombia. These community nurseries are part of a national program, which has been implemented about 40 years ago and targets families from the lowest stratum. Bernal, Fernández, Flórez, Gaviria, et al. (2009) evaluated the program and reported that many caregivers have insufficient knowledge about infant development. Therefore, a local NGO implemented an intervention with four components to improve the quality of care. First, community mothers received a formal vocational training as early childhood teacher. Second, they received support and coaching to integrate what they learned in their daily nursery routines. Third, they learned how to teach parents about issues related to appropriate child care and development. Fourth, the implement- ing NGO monitored and supported the children in school, once they left the community mother. While the existing research has focused on cognitive and socioemotional skills, I complement the analysis with an elaboration of the effect on individual preference mea- sures. For that, I conducted lab-in-the-field experiments with children who have visited a treated nursery and compare them to children who have been in an untreated nursery. I elicit measures for prosocial behavior, egalitarian preferences, trust, risk and time pref- erences. Moreover, I do evaluate not only the effect in the short run but also four years after the implementation. The intervention improved children’s cognitive, psychosocial, and psychomotor skills. More importantly, the effects persist up to four years after the intervention. We show that children from the treated group have better grades and are more likely to be in the right grade for their age. Interestingly, the intervention also af- fected social preferences of the children. Children who visited a treated nursery are more altruistic than children who attended a regular nursery. The evaluation set-up does not allow to fully rule out that unobserved factors drive the effects. Furthermore, we are not able to identify which part of the intervention is responsible for the observed effects. But the pattern of results makes me confident that what I see can be traced to the intervention.

Create date
06/10/2017 10:40
Last modification date
20/08/2019 14:08
Usage data