Researchers may deceive themselves when they mislead study participants

By Bruce Bower

Web edition: October 22, 2010

BASEL, Switzerland — As dusk settled over this charming city by the Rhine in early October, psychologist Ralph Hertwig sipped scotch in his office with a visiting journalist and bemoaned the toxic — and for some researchers, intoxicating — effects of telling lies to gather data and get published.

Hertwig’s theme: Inauthentic experimenters and the research subjects who follow their lead. His case in point: A study in the May Psychological Science reporting that people who wear discount, mock designer sunglasses feel phony as a result and become more likely to cheat and to judge others as unethical.

With apologies to Jerry Lee Lewis, there was a whole lotta fakin’ going on in this investigation. Half of female participants in one trial completed a bogus questionnaire and were told that their answers reflected a preference for counterfeit products. They were then instructed to take a pair of sunglasses from a box marked “Counterfeit Sunglasses” — which actually contained expensive designer shades — and wear them while walking outside the lab for five minutes and then while working on lab tasks that paid money for correct responses.

Volunteers recorded their responses on a work sheet, after having been promised anonymity by the experimenters. But numbers on work sheets were used to identify each responder so that her actual and self-reported performance could be compared.

And behold—relative to women who hadn’t been misled about favoring faux stuff, tricked participants claimed to have made more correct responses than they actually did. In another experiment, misled women frequently described others as unethical and devious.

Hertwig rubbed his eyes wearily. “It’s just as likely that the experimenters’ own behavior encouraged the dishonest behavior that they observed,” he said.

Participants in the counterfeit condition could have read the situation as one in which normal standards of behavior didn’t apply because the researchers approved of designer knock-offs, Hertwig explained. Each woman saw that the experimenter had somehow acquired fake designer gear and displayed it openly. What’s more, the experimenter claimed special insights into people’s likings for counterfeit products, told volunteers to wear the glasses in public and had them evaluate positive statements about the glasses.

Sociologists’ “broken-windows” theory posits that signs of disorder and petty criminal behavior cause such acts to spread in communities. If that’s the case, Hertwig noted, the counterfeit-sunglasses scientists metaphorically “broke their lab’s window and cried foul when participants sprayed graffiti on the wall.”

And assuming volunteers were debriefed after the experiment, as required by the American Psychological Association’s rules of conduct, one shouldn’t expect them to trust any future researchers’ pledges of anonymity.

Ironically, psychologists’ blindness to these issues could stem from a counterfeit-sunglasses effect. “Deceptive research practices may induce a sense of self-alienation and lack of authenticity among experimenters that interferes with analyzing the signals that the experimental situation conveys to participants,” Hertwig mused.

Some of the most famous psychology experiments of the past 60 years have hinged on trickery, despite longstanding ethical and practical concerns about fooling people in the name of science (SN: 6/20/98, p. 394).

Deceptive psychology’s heyday occurred in the 1960s and 1970s. Literature searches conducted by Hertwig and economist Andreas Ortmann of the University of New South Wales in Sydney, Australia, indicate that experimenters still mislead volunteers in between one-third and one-half of studies published in major social psychology journals.

Hertwig doesn’t want to ban deceptive research practices. He’d settle for researchers taking off their rose-colored, counterfeit sunglasses and scrutinizing how their devious methods may shape volunteers’ responses.

In other words, let the liar beware.