Insurance Demand Experiments: Comparing Crowdworking to the Lab

We analyze insurance demand experiments conducted in two different settings: in-person at a university laboratory and online using Amazon's Mechanical Turk crowdworking platform ("MTurk"). Subject demographics differ significantly across the samples, but average insurance demand is similar. However, MTurk subjects have more variation in their demand and react less to changes in exogenous factors of the insurance situation. Applying filters to the MTurk population does not lead to more comparable demand between both samples. Additionally, while MTurk subjects pass comprehension questions at the same rate as in-person subjects, they show more random behavior in other questions. Based on our experiment, we outline the advantages and disadvantages of each platform, suggest some best practices for conducting insurance demand experiments using MTurk, and draw implications for future studies using this subject population.