An in-class exercise in a research methods class illustrated the process of operationalising a construct in a policy-relevant context that was familiar to college students. Students read about the drunk driving problem, and were asked to estimate the number of beers needed to push their blood alcohol content (BAC) to .10 percent, the level in most states where a driver is guilty of drunk driving. Results showed wide variation in students' estimates, even after controlling for gender and weight differences. This variation demonstrates how uncertainty may continue to surround operationalisation, even when the chosen indicator is intended as a policy improvement. The exercise is potentially applicable to statistics or social problems courses, and those applications are discussed. (Author/publisher)
Abstract