<< A backhanded compliment? | Main | Cal. AG Debate >>

Methodology and Lying

In social science research, there are always variables you can't control, so you try to structure your study so that the things you can't control are random with respect to the variables you are trying to measure. Randomness plus large enough sample size plus a lucky rabbit's foot means these uncontrollables don't affect your results. You hope.

BBC reports on a British Science Museum study where the data are gathered through a survey. That is common in research, but we always have to worry that survey respondents sometimes lie to surveyors.

What is the variable of interest? Lying.*

So how do the researchers know their results aren't skewed by lying participants.**

The BBC story doesn't say, and I can't find a link to the full study. Maybe they have some clever way. They certainly can't assume that lying is random with respect to the variable of interest, though.

The researchers might say, "Nothing's wrong, I'm fine." That is the number one lie told by women and number two for men. If the participants are telling the truth about not telling the truth, that is.

* A variable of keen interest to criminal trial lawyers, which is why this isn't entirely off topic.

** We don't say "research subjects" any more. That will get you cited by the Language Police.


I might be a fool for asking, but what's the No. 1 lie told by men?

I didn't have that much to drink.

I thought it was going to be, "No, you don't look fat in that dress."

Leave a comment

Monthly Archives