Study reveals that students aren’t ideal test subjects for behavioral research
Students are popular test subjects for many studies in behavioral sciences. However, using only students does not reveal the full picture about people in general. In fact, many of the students’ decisions in those experiments differ from those of other population groups. These are the findings of a new, extensive study consisting of 36 experiments which was conducted by a team of behavioral scientists from Martin Luther University Halle-Wittenberg (MLU). The study was published in the Australian Journal of Agricultural and Resource Economics.
Students are popular test subjects, especially in behavioral economics. “This makes sense because students are open to and interested in scientific studies, they are already in a university setting, and also they are receptive to financial incentives offered in those studies,” says Dr. Sven Grüner from the Institute of Agricultural and Food Sciences at MLU. “However, it is unclear whether students are representative of other population groups—after all, they differ in important ways, such as age and income.”
To answer this question, the behavioral economist conducted an elaborate study using 300 test subjects. The result: Only limited conclusions can be drawn from students about the behavior of other people.
In a total of 36 sub-experiments, Grüner compared the decisions of agricultural science students to those of farmers. The team examined individual characteristics such as risk-taking, impatience, altruism, trust, punishing unfair and rewarding generous behavior. “We used established economic experiments from decision and game theory,” Grüner explains.
For example, when determining willingness to take risks, the test subjects were given the choice between a higher probability of winning a small amount of money and a lower probability of winning a higher amount of money. “We gradually increased the monetary incentives in all of the experiments to see how the expected sum influenced decisions,” says Grüner. Contrary to previous studies, the incentive was actually given to the subjects afterwards, because theoretical earnings could falsify the results: If participants knew they would not get any money, they might have shown a greater degree of socially desirable behavior.
The results of the comprehensive study revealed a very mixed picture: for example, there are no clear differences between the groups in terms of risk-taking. “This contradicts earlier studies in which students were more risk averse than farmers,” says Grüner. The differences were also slight when it came to trust and rewarding generous behavior.
However, when testing for the groups’ patience there were bigger differences: farmers were much more likely to choose the option with the higher probability of a lower pay out, while students were consistently found to be more patient and wait longer for more money. At the same time, farmers turned down unfair offers more often, even when this meant they would not get any money themselves. These finding are not in line with earlier studies that showed similar behavior among students and other population groups, says Grüner.
“Our study shows that it is really problematic to generalize the behavior of students to other real actors. This could call into question a lot of the results of previous studies—not only in agricultural sciences, but across all disciplines,” adds Grüner.
This is also a sensitive topic because surveys on individual decision-making examine important questions about the future—risk behavior and patience, for example, are decisive criteria for investing in sustainable production structures that usually only pay off after many years. The new study helps to identify factors that can be used to weight the results.
Sven Grüner et al, How (un)informative are experiments with students for other social groups? A study of agricultural students and farmers, Australian Journal of Agricultural and Resource Economics (2022). DOI: 10.1111/1467-8489.12485
Study reveals that students aren’t ideal test subjects for behavioral research (2022, September 12)