Dewey Wins Again! UK Edition…..

Because nothing is more important than made up numbers.

Because nothing is more important than made up numbers.

Well, Ralph Miliband’s boring son and the Labour Party lost big across the pond last week, and the crushing defeat was completely unexpected. Shit polls and pollster hacks like Nate Silver were quite confident that Labour would rule the day and all of Brittania….Only they lost. Big. Just like Eric Cantor. And Dewey.

This is a function of the worldwide decline of scientific polling, and the growth of polling profiteers. The idea that we should have daily polls supposedly tracking “changing opinions” based on whether a politician farted or something is ridiculous. All this has done is to misinform parties, candidates, the media, and citizens about the opinions and preferences of the public. Aggregating bad data doesn’t help, nor does it make any sense to make data aggregators into revered shamans.

The decline in data quality and scientific polling has also led to the disturbing trend of using Cheetos munching weirdos from marketing panels in all kinds of “research.” These panels are not drawn from a random sample of the population, and panelists are expected to complete at least 4 surveys a year in or else they’ll only get baked Cheetos or be “withdrawn.” No attempt is made to assess the veracity of the data completed on-line by who knows whom. We saw this, of course, in the Regnerus fiasco, and the severity of the problems with this poll received more attention in a new reanalysis of his shit data.

But, the problem with reanalyzing shit data is that stirring the shit just makes it smell worse. Nobody should ever be allowed to publish a paper in a journal with “science” or “sociology” in the title analyzing data from a shit poll. It isn’t just Regnerus. GLBT activist scholars are busy right this very moment analyzing a shit poll taken from the exact same panel used by Regnerus. And, that 7 foot 8 guy who weighs 88 pounds is probably picking up his case of Cheetos for that one as well.

What kind of weirdo would agree to fill out longwinded polls online in exchange for Cheetos? How many Americans are really that connected to the internet, even? Are ethnic minorities who are online just like other ethnic minorities? What about the elderly? Can we just assume that the non-randomly recruited 80 year-olds in some online marketing panel are just the same as other old people? Yeah, just adjust the data. Right.

In the last year I’ve reviewed four or five papers using data from non-random., online marketing panels. If you are a sociologist and you’re using such data, stop. If you continue, you are hurting scholarship and likely hurting your career and those of your students who are being encouraged to take short cuts with dirty unscientific data.

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s


%d bloggers like this: