Monday, April 18, 2011

The unreliability of sexual orientation polls

| »

The Wall Street Journal has a detailed report on the difficulties of establishing just how many gay, lesbian and bisexual people are present in the US. To give you an idea, here’s a little graph illustrating the vast discrepancies in the results of various polls, depending on how they’re phrased and presented:

Graph: “Identity Gap: The percentage of American adults who tell researchers they are gay, lesbian or bisexual varies depending on how the question is phrased.” Results: 1.7% (National Epidemiological Survey on Alcohol and Related Conditions, 2004-05, asked in person), 5.6% (National Survey of Sexual Health and Behavior, 2009, asked online), 3.0% (National Election Pool exit poll, 2010, asked by paper questionnaire or phone)

It’s obvious that asking people in person whether or not they’re homosexual tends to make them slightly more guarded, especially in environments where they may not be free to speak freely, but it’s nonetheless troubling – albeit unsurprising – to see such a drastic variation depending solely on how the poll is asked. Polls like these are what’s used to determine LGBT policies and can drastically influence local and national politics, so it’s probably high time pollsters settled on a single, clear and all-inclusive method of determining just how many LGBT folks there are in the country.

Of course, the answer is simple: Use the Kinsey scale (which, despite its criticisms, is still a remarkably simple yet effective way to pinpoint just how gay people are(n’t)), and ask all such questions by phone or through paper or online questionnaires (trying to avoid online polls as much as possible), and always anonymously, to ensure that no-one feels the needs to misidentify themselves based on their environment.

I know, I’m a (predominantly heterosexual, only incidentally homosexual) genius.

(via Joe. My. God.)