You learn things as a blogger from your readers. I once wondered what is going on with those cheesy ads you see on a lot of mainstream media websites, usually down near the bottom—ads like “you won’t believe what this celebrity looks like today!”, “10 of the worst wedding-day fashion mistakes!”, “Amazon hates when its customers know this secret”, “Do this one thing and you’ll be happier and wealthier!” and so on.
Several readers responded, “They’re called chumbuckets or chumboxes” and it turns out they’re one of the consistent money-makers across most of the history of the World Wide Web. I assumed the opposite, that nobody ever clicks on them, partly because those of us who were there at the dawn of the web learned the hard way never ever to click on an active link in a webpage where we didn’t know in advance what we’d end up at. Having a link open a thousand webpages a second just to register clickthroughs, leading to a frantic control-alt-del to shut the whole computer down, teaches you some caution.
But you’d think also that “do this one thing!” and most of the other appeals of the chumboxes would make the average user wary anyway. Unfortunately not: a sucker born every minute. The impulse to click on that kind of appeal is more widely distributed than you might think, in fact. I think it’s the single impulse behind many of the most prominent examples of manipulated or fraudulent research in academic communities.
Andrew Gelman, writing for the Chronicle of Higher Education, offers a kind of meandering take on a number of prominent examples of fraud in psychology and behavioral economics, including the recent case of Harvard Business School professor Francesca Gino’s research on dishonesty. Gelman’s take is a mix of trying to guess at the mindset of prominent scholars who manipulate data or fudge experiments, a bewilderment at why researchers studying dishonesty and cheating aren’t a little more diligent about data supplied by collaborators given the subject matter, and a very gentle suggestion that maybe some disciplines or some types of research allow for faking data more easily.
I’m going to be more direct about this: cheating of this kind is almost entirely a problem of non-qualitative social sciences. Not indiscriminately so: almost invariably this kind of research fraud is associated with scholarship that claims this one thing is a major, easily manipulated variable in producing what otherwise seems to be complex individual and collective behavior.