Friday, July 06, 2007

quiz feature!

Received wisdom among those who run survey centers is that some interviewers are better than others at convincing respondents to participate in surveys. This is easier to assert than actually show. As some readers know, I am an investigator on the Wisconsin Longitudinal Study, a survey of thousands of 1957 Wisconsin high school graduates that fielded its last round of phone surveys ~3 years ago. After several hours of meticulous data analysis, I have adduced strong evidence indicating that, in fact, some WLS interviewers were better than others. Today is thus not a day that feels much like I am splitting the sociological atom, here.

WLS respondents were in their late sixties; WLS telephone interviewers were mostly undergraduates (or undergraduate-aged). Any guesses about whether there was a sex of interviewer effect in the WLS (that is, a tendency for male or female interviewers to be better at getting respondents to participate)?


Ken Houghton said...

Ask an easy one next time. Female surveyers likely got a better response rate. Even with the WLS demographic skew, I'd spot you 5-10% on the differential before talking seriously about a betting line.

B said...

I used to work for the Survey Center and called on WLS. Being male, I found that I could convince female SOs of the grads to participate much more frequently than male SOs. It was semi-rare for me to have any difficulty with an actual grad (though when there was it seems like they were always male), it was always the SOs that wanted to refuse.

On a side note, calling for WLS was one of the few studies that was usually fun to call on (lots of participants seem to actually enjoy it rather than feel bothered) and I would actually look forward to work.

jeremy said...

B: A former WLS interviewer reads my blog? Really? BTW, the better the WLS primary respondent reported getting along either either their spouse or their sibling, the less likely the spouse or sibling was to decline to participate. Not surprising, but interesting to think of what it's implications are for what kinds of couples and sibships are disproportionately represented in the data.

B said...

Yep :P I worked for the survey center for about a year, a couple years back (it was around the time they were starting the wave of SO surveys until shortly after they started the sibling surveys). I can't claim to have been reading your blog during that time as I just started reading several months back, though.

eszter said...

I checked on something similar recently in my data set. Three people conducted interviews and then tried to get people to participate in the text-messaging follow-up study. If I recall, one of the three did about twice as well in recruiting into the follow-up study than the others. All were female.

shakha said...

I'm going to guess that men have an easier time. Why?

1.) It's easier to say "no" to a woman!
2.) The men are taken as being "more scientific". So answering their questions is seen as less of a waste of time. (my REAL explanation).

Which reminds me of that old experiment where students were asked to evaluate the quality of a study. The manipulation was changing the sex of the author. Sure enough, people thought men knew the field better and gave a more favorable evaluation. Women had an even stronger bias than men. The weird thing about the experiment was that there was no guarantee that subjects looked at the author's name. But that's kinda irrelevant.

So my claim: MEN are better!

shakha said...

so... what's the answer?!?