Tuesday, February 19, 2013

Of Food Banks and Surveys

I often point out surveys sponsored by special interest groups that happen to find people don't know enough about, coincidentally, that special interest's, um, interest.  Beware such self-serving polls, even if they're for a good cause, like this one out of Texas by food banks that finds 1-out-of-2 north Texans know someone who has used a food bank.

Again, good cause. It even has a pdf with more graphical display of data. I'm surprised that 80 percent of respondents say they've contributed in some way to a food bank.  That seems high, but if true -- all the better.

Now, to methodology.  They report:
There were more than 1,600 respondents to the survey, which was conducted in January 2013. The survey was conducted via an opt-in email targeting more than 60,000 North Texans, as well as via Facebook. 
You can't tell a lot from this, other than it's a hefty sample size.  What I'm concerned about, obviously, is how well these 60,000 folks represent the population as a whole.  It's unclear where the 60,000 names came from.  If from the food bank's own list, or lists they have access to, then you have to worry about the generalizability of the results.  There simply isn't enough methodological information to make a judgment -- and when that happens, you're spidey sense should go off.

No comments: