Lab vs Online Experiments : Gender Differences

Comparing the behavior of subjects randomly allocated to perform tasks and tests online or in the lab, I find that there are differences that vary according to subjects’ gender. Selfishness is activated by the online environment, where the level of observation is lower, for males only; only females are more risk-averse when they answer online. This evidence highlights how the study of gender differences can be biased by experimental conditions. Interestingly, females’ performance in math tasks is influenced by the administration conditions: online they perform better. This corroborates the hypothesis that the way used to implement tests can affect the gender gap in educational achievement. As online environments have become more common, both for experiments and educational purposes, it is important to understand how this shift affects gender gaps. This work provides preliminary results about how gender and administration conditions interact