Stupid research

So today’s copy of the Journal Inquirer (JI) has the provocative headline: Students mine data to find where unfaithful husbands live: Fairfield County. Well it’s actually a Toledo Blade story — the Connecticut based JI just tacked the “Fairfield County” on the end.

So these kids at University of Toledo use data from hacked Ashley Madison website to determine who cheats most often, noting that “Metropolitan areas with some of the lowest subscription rates were in poor Appalachian and southern locations, strengthening the conclusion that affluence is linked to this kind of online adultery.” The story goes on to note that “The Fairfield County area also topped the list of metro areas in spending rates, doling out $1,127 on the site per 1,000 people.”

Huh. Of course this fits with the usual liberal narrative that rich people are inherently evil and poor people inherently virtuous ….

One thousand, one hundred twenty dollars ? And that’s just on the website — presumably there are other costs in cheating. Or maybe — just maybe — poor people have cheaper ways to cheat. Because the Internet is a tiny place and Ashley Madison is the only way to cheat …



One thought on “Stupid research”

  1. Here’s the actual paper:

    They do a decent overview of different factors that researchers think influence cheating. Affluence was not the only one. In this paper, they try to use the Ashley Madison dataset to evaluate which factors have the highest correlations with being a user.

    The paper is much more balanced and dispassionate than the two articles you linked to. They could have done a better job of accounting for confounding effects. As an example, my wife and I subscribe to Sun Basket, a delivery service that, every week, sends ingredients and recipes for making 3 dinners. I would argue there is probably a correlation between income and luxury online services in general. A couple with a lower income might just use a cookbook to find new meal ideas.

    But, I have faith that another researcher will come along, identify the flaws, and publish a follow up, which improves the analysis by accounting for things the original authors didn’t.

    There is, of course, the issue of whether this research is a worthwhile usage of resources. I would bet that the students did this in their spare time, not funded by a grant or as a part of a professor’s project. Either way, I imagine the utility is up for debate, and we may not find agreement there.

    The “news” articles are absolute trash, though. I agree with there. They picked out correlation, misinterpreted it, and spun it. It’s pure clickbait and misrepresentation — designed to fit a narrative that drives clicks. This is not uncommon when “journalists” try to popularize science. We need better science reporting — it just makes decent work look bad.

Leave a Reply

Your email address will not be published. Required fields are marked *