How (un)informative are experiments with students for other social groups? A study of agricultural students and farmers

Experiments are often used to study individual decision-making under controlled circumstances. Due to their low opportunity costs and high availability, university students are frequently recruited as the study population. Even though they are rather untypical with regard to many characteristics (e.g. age and income) compared to the representatives of the social group of interest, the experimental behaviours of students are sometimes prematurely generalised to other social groups or even to humans in general. Given the widespread challenges in the agricultural and environmental sector, it is particularly interesting to address farmers' decision-making. We analyse whether agricultural students can be used to approximate the behaviour of farmers in simple economic experiments, which are often used to measure risk aversion, impatience, positive reciprocity, negative reciprocity, altruism and trust. Moreover, we consider the role of systematically varied monetary incentives. We find no differences between agricultural students and farmers in their risk aversion; farmers' positive reciprocity and trust are positively associated with the incentive level, which cannot be observed with agricultural students. Findings regarding altruism in the two populations are mixed and challenge the finding of earlier studies of students being less pro-social. Agricultural students are a lower boundary of impatience and negative reciprocity. These heterogeneous results suggest that scientific inference from agricultural students to farmers should be made cautiously. However, we do not deal with a representative sample of our target population (e.g. gender). Replication studies are required to evaluate the generalisability of our findings.