Question

Suppose that you wish to find out what actions people are actually taking to prevent the spread of AIDS and other sexually transmitted diseases. You obtain a brochure from the Centers for Disease Control that indicates sensible precautions to take, and then you use this to design a survey to examine whether or not people actually take these precautions. What kinds of problems are you likely to run into in administering such a survey? What issues of social desirability bias should we worry about? Discuss anonymity of findings, social desirability scales, and/or unobtrusive measures (for example, condom sales: if students claim to be practicing safe sex but condom sales in the area are low, then you would suspect that students were answering your questions in a socially desirable way) as possible means of reducing social desirability concerns.

Answer 1

There are several problems that you could run into when administering such a survey. The most obvious one would be social desirability bias. This refers to a type of response bias in which respondents answer questions in a way that will make them be seen favourably by others. This could be a problem in this research, as people would most likely want to appear responsible, and they might underreport dangerous behaviour. This would also be problematic as certain social groups (ex. teenagers, women) could be even more underreported than others.

One way to combat this would be through anonymity and confidentiality. If respondents are able to answer without their answers being linked to them, they would be more likely to answer honestly.

Another strategy could be using unobtrusive measures. This means that data collection does not impinge on the subjects of the study, and that the researcher makes no direct contact with the subject of the study. However, this would present its own problems as well. In particular, this would be difficult when it comes to medical records, as these are considered confidential.