Research

I do research to deepen our understanding of human behavior spanning childhood, adolescence, and later stages within organizational and labor market contexts. Importantly, the body of existing evidence, including my own contributions, indicates that individuals display a wide array of responses to institutions and incentives, driven by differences in personal attributes such as gender, economic preferences, attitudes and proneness to cognitive biases. I firmly believe that gaining a deeper understanding of these attributes and their influence on behavior across diverse contexts can yield valuable insights. Ideally, these insights would result in more precisely targeted policy interventions and the development of more equitable and efficient institutional structures. 


Publications

Competition and prosociality – a field experiment in Ghana

Abstract: 

Competitive bonuses are commonly used to promote higher productivity in the workplace. Yet, these types of incentives could have subsequent negative spillovers on coworkers' prosocial behavior. We extend this line of research by comparing the effect of competitive bonuses on attitudes towards others  (measured using the social value orientation) and changes in the propensity to cooperate in a public goods game (PGG).  Using a lab-in-the-field experiment, we find that compared with threshold payments, competition reduces prosociality (as measured by both SVO and PGG) when the dispersion of payments is high.  However, when there is less at stake, competition does not affect prosociality.  The main mechanism seems to be associated with rivalry that is created toward others and a feeling of entitlement among high-income earners. We find that while attitudes towards others are lower in competition compared with random payments, there are no differences in cooperation. This suggests that competition has a distinguishable effect on prosociality that can be associated with rivalry. 

Citation: Grosch, K., Ibañez, M., & Viceisza, A. (2022). Competition and prosociality: A lab-in-the-field experiment in Ghana. Journal of Behavioral and Experimental Economics, 99, 101887. 

(citation download is available here)

Paper     Instructions     Data and Do-File


Unfair wage differentials and their effect on antisocial behavior

Abstract: 

In this paper, we investigate how payment procedures that are deemed unfair can spur unethical behavior toward innocent co-workers. More precisely, we analyze how unfair wage differentials affect antisocial behavior in a real-effort experiment. In our Discrimination treatment, a highly unfair payment procedure with wage differentials, half the workforce is randomly selected and paid by relative performance, whereas the remaining receives no payment. A Joy-of-Destruction game measures unethical behavior subsequently. In control treatments, payments are completely based on relative performance or completely random. Non-earners in Discrimination destroy significantly more than in the controls. In Discrimination, unethical behavior is generally high for all non-earners, independently of individual inequality aversion and their relative performance beliefs. By contrast, inequality aversion is the main driver of destructive behavior in the control treatments. 

Citation: Grosch, K., & Rau, H. A. (2020). Procedural unfair wage differentials and their effects on unethical behavior. Economic Inquiry, 58(4), 1689-1706.

Paper


Gender differences in honesty – the role of social value orientation 

Abstract: 

This paper experimentally analyzes the role of individual social value orientation (SVO) on honest behavior. We focus on a situation where dishonest behavior pays off at somebody else's cost. In this case, distributional preferences might matter for the willingness to act honestly. To examine this link we conduct a laboratory experiment where we first elicit SVO to measure distributional preferences. Afterward, we implement a die-rolling game to elicit dishonest behavior at an individual level. We detect a positive correlation between subjects' SVO angle and honest behavior. Furthermore, the data confirm common gender differences, i.e., women are significantly more honest than men. Additionally, we find that, on average, women have higher SVO angles than men. A mediation analysis reveals that SVO explains the gender effect.

Citation: Grosch, K., & Rau, H. A. (2017). Gender differences in honesty: The role of social value orientation. Journal of Economic Psychology, 62, 258-267.

Paper     Instructions     Data and Do-File     z-tree


Other publications


Grosch, Kerstin and Rau, Holger A. (2021) Gender differences in leader's compliant behaviour. An interview with Kerstin Grosch and Holger A. Rau. CEJ Compliance Elliance Journal, 7 (2), pp. 2-5.

Grosch, Kerstin; Müller, Stephan; Rau, Holger A. and Zhurakhovska, Lilia (2021) Gender differences in leader’s compliant behaviour – Selection into leadership and dishonest behaviour of leaders: a gender experiment. CEJ Compliance Elliance Journal, 7 (2), pp. 6-11.


Working Paper

The school guide - expert insights and best practices for the effective implementation of school experiments

Abstract: 

This guide provides a comprehensive overview of the distinct characteristics of school experiments conducted with children in preschools and schools. We investigate and describe the essential considerations involved in designing and implementing such experiments, drawing insights from a survey conducted with senior researchers. Moreover, the guide summarizes nine key lessons learned from the experiences of these researchers. The paper also presents the opinions of inexperienced researchers in school experiments (juniors) on crucial aspects of successful school experiments, which differ from the opinions of experienced senior researchers. As a result, this guide serves as a valuable resource for junior researchers embarking on their first school experiments. Promoting the adoption of best practices endorsed by senior researchers, strengthens the validity and reliability of school experiments. 

Paper (as of June 2023)

Suggested citation: Grosch, Kerstin & Haeckl, Simone & Rau, Holger & Preuss, Paul, 2023. "A Guide to Conducting School Experiments: Expert Insights and Best Practices for Effective Implementation," UiS Working Papers in Economics and Finance 2023/2, University of Stavanger 

(citation download is available here)


Closing the gender STEM gap - A large-scale randomized-controlled trial in elementary schools

Abstract: 

We examine individual-level determinants of interest in STEM and analyze whether a digital web application for elementary-school children can increase children's interest in STEM with a specific focus on narrowing the gender gap. Coupling a randomized-controlled trial with experimental lab and survey data, we analyze the effect of the digital intervention and shed light on the mechanisms. We confirm the hypothesis that girls demonstrate a lower overall interest in STEM than boys. Moreover, girls are less competitive and exhibit less pronounced math confidence than boys at the baseline. Our treatment increases girls' interest in STEM and decreases the gender gap via an increase in STEM confidence. Our findings suggest that an easy-to-implement digital intervention has the potential to foster gender equality for young children and can potentially contribute to a reduction of gender inequalities in the labor market such as occupational sorting and the gender wage gap later in life.  

Paper (as of August 2022) Online Appendix (including instructions)


Contract compliance with overconfident expectations: Experimental evidence on reference-dependent preferences

Abstract: 

This study examines the effect of agents' overconfident expectations in their production on their contract compliance. Drawing on a reference-dependent framework, we theoretically derive implications for situations where an agent exhibits overconfidence and loss aversion. We further conduct an experiment with students from Ghana, a country characterized by limited external contract enforcement. We find that overconfident agents comply at the same rates as unbiased agents. However, highly overconfident and loss-averse agents breach to a higher extent. In a treatment, we manipulate agents' overestimation exogenously and use it as an instrument to establish causality.

Paper (as of April 2024)     Online Appendix


Selection into Leadership and Dishonest Behavior of Leaders: A Gender Experiment

Abstract: 

Leaders often have to weigh ethical against monetary consequences. Such situations may evoke psychological costs from being dishonest and dismissing higher monetary benefits for others. In a within-subjects experiment, we analyze such a dilemma. We first measure individual dishonest behavior when subjects report the outcome of a die roll that determines their payoffs. Subsequently, they act as leaders and report payoffs for a group including themselves. In our main treatment, subjects can apply for leadership, whereas in the control treatment, we assign leadership randomly. Results reveal that women behave more dishonestly as leaders while men behave similarly in both the individual and the group decision. For female leaders, we find that sorting into leadership is not related to individual honesty preferences. In the control, we find that female leaders do not increase dishonesty. A follow-up study reveals that female leaders become more dishonest after assuming leadership, as they align dishonest behavior with their belief on group members’ honesty preferences.

Paper (as of April 2021) 


Gender equivalence in overconfidence - A large-scale experimental study in a non-WEIRD country

Abstract: 

This study examines gender differences in overconfidence, focusing on overestimation (individuals' perception of their performance relative to their actual performance) and overplacement (individuals' perception relative to the performance perception of others).  We conduct large-scale lab experiments with over 1000 participants in Ghana, a non-WEIRD country in Sub-Saharan Africa. It examines overestimation and overplacement measured in an incentivized way and employs various analytical approaches to examine gender differences. Unlike previous studies that found men to be more overconfident in a task where men outperform women and subjects also anticipate the gender difference in performance (male-typed tasks), this study reveals no significant gender differences in overestimation in such a task. Concerning overplacement, women and men do also not differ significantly in their belief to be better or worse than other subjects of their own gender.  Moreover, individuals who overestimate their performance are also more likely to believe they outperformed others, regardless of gender. Overall, results indicate gender equivalence in overconfidence. Gender differences in overplacement only arise when comparing individual performance estimates to those of the opposite gender which persists after accounting for performance differences, with women more inclined to view themselves as inferior to men.

Short Paper (as of May 2024)     Instructions



Work in progress

Peer effects in adolescents' saving behavior

Abstract: 

In real life, the saving behavior of peers is much more difficult to observe than consumption behavior. What does it mean to make savings behavior of peers also visible? How do peers influence people's savings behavior? These are questions that we examine by using behavioral economics theories and experimental methods. Our study focuses on adolescents. It is important that young people learn how to manage their money to be protected for (potentially) future economic shocks. Behavior and preferences are shaped early on and are still malleable at a young age. Moreover, we focus on young people because they are about to earn a self-managed income and, thanks to the class/school association, natural surroundings of peers are available for the investigation of our research questions. 


The hidden cost of violent conflict - Sorting into Local Labor Markets in Colombia

Abstract: 

Violent conflicts have negative effects on prosperity and development. Reconstruction efforts require that a qualified labor force be willing to work in highly violent areas.  We used a field experiment to investigate the effects of life risk on sorting in the labor market. We offer comparable jobs in low- and high-conflict areas in Colombia to a pool of job seekers. Application rates decrease by 12~percentage points due to riskiness.  An increase in salary helps increase the application rates for high-risk jobs but does not close the gap. We found no evidence of differential effects for better-qualified applicants, early childhood exposure to violence, or gender on the probability of applying for a job in a district of high or low violence.


Further ongoing projects



Recently (planned) field experiments without an academic publication

Incentivizing citizens’ environmentally-friendly mobility behavior – a field experiment in Austria 


Nudging to reduce littering in waste disposal areas – a field experiment in Austria’s community houses