Diederik Stapel. Brian Wansink. Nicolas Guéguen. Anyone who's been following recent debates over research integrity in psychology will recognize these as three prolific and successful academic psychologists who have suffered a total (Stapel) or ongoing (Wansink, Guéguen) fall from grace in the past few years. If you're not familiar with these cases, you can start by reading over Nick Brown's blog. Brown has been at the centre of the investigations into irregularities in Wansink and Guéguen, and he also translated Stapel's book into English. In this post, I will not revisit the issues with these three psychologists' work. Rather, I want to point out that, taken at face value, their work was brilliant. They published research that, if real, would have been a model of what experimental psychology should be.
Stapel, Wansink and Guéguen - let's call them SWaG - worked on different topics, but what all of them had in common was that they published naturalistic experiments. Conventional psychologists carry out experiments, but not naturalistic ones. They invite people (students mostly) to come into a testing room and perform some tasks or assessments in return for money or course credit. It's an artificial environment. The participants know they are being studied. It's not natural, not real-life. SWaG's experiments, by contrast, were realistic (if not necessarily real). Guéguen, for instance, reported sending a female investigator into real bars to see how long it took for real men to hit on her, depending on the heels she was wearing, which varied. A silly topic perhaps, and the study may have been fabricated, but setting those issues aside... it's great. It's an application of the experimental method to actual human behaviour in 'the wild'. I don't think that a conventional lab psychology experiment would be able to tell us whether high heels help women to get hit on. Lab studies are well-suited for the study of perceptual or cognitive processes, but for complex, social behaviours, I'm not convinced that the lab is the right place to experiment. My point is, maybe psychologists should look at SWaG and "do as they say, not as they do". Don't do what they did, but try to actually do the kind of things that SWaG claimed to have done. Of course, it's easy enough to say that experiments should be naturalistic. Naturalistic validity is difficult to achieve in practice. Perhaps the most fundamental problem is ethics: by definition, a truly naturalistic study would involve experimenting on people without their consent, because if they knew they were consenting to take part in a research project, it would no longer be a real-life situation. What ethics review board would approve such a project? It is unclear whether Guéguen got official ethical approval for his studies - although if these studies never actually happened, I suppose he didn't need it. But a legitimate naturalistic experimenter would need it, in most parts of the world (although maybe not in France, where Guéguen works.) Perhaps there is a lesson here for our ethical standards. The fact that SWaG's work was highly cited shows that naturalistic work is valued by the research community, and - as far as I know - there was no public outcry over the ethics of these kinds of experiments, at least not until the scandals started. Yet it seems that it is very hard to do these kinds of studies without resorting to SWaG-like shortcuts. Perhaps we need a more open-minded approach to the ethics of naturalistic experiments?