What I really like about your article Tom is that the last bit is an ace attack on positivism.

The reporting of psychology stuff in the press is always positivist, and that is what gives it its magical flavour. It all subtly promotes the idea that the ideas in our heads are kinda illusions. We may think that we do things because of stuff that we are thinking or feeling, but REALLY we are being governed by forces beyond our control, which special authority figures called scientists can reveal to us. It is thus an attack our dignity and autonomy, and on the view of us as reasonable creatures who can be talked to as well as controlled with prodding etc (and thus many other things, including – I think – a subtle attack on democracy).

The second you relate drinking speed to people’s ability to estimate how full a glass is – to a theory about what people believe – you pull it back from that kind of magical positivism, and it slots back into the normal folk psychology that is compatible with our being creatures of reason, meaning and dignity.

The idea for this study came out of a conversation between myself (I research tobacco and alcohol use, mainly, from a number of perspectives) and a vision scientist (with an interest in psychophysics). Not surprisingly, it was over a drink, and we were wondering why there seemed to be such an increase in the number of odd-shaped, branded glasses in which beers were being served. I think it’s quite a nice example of how people from different research backgrounds can come together to try to begin to answer complex questions.

Of course there are lots of limitations (not least the sample – psychology has been described as the science of psychology undergraduate students), and I would never want to over-state the results. There’s also perhaps an interesting discussion to be had about whether we (consciously or unconsciously) over-stated our findings because of the current impact agenda in UK research. On the other hand, there is at least a possibility that our results will translate to the real world. We’re obviously planning follow-up studies.

In the UK one of the strongest determinants of drinking rate is whether you are drinking in a round. Drinking alone, with few external influences, helps to isolate specific effects, but it also means the results can’t easily be generalised. So we have to strike a balance between experimental purity and generalisability. As a first step, we generally err on the side of the former, but ultimately we need to get to the latter. I think psychology often does a pretty poor job of this and stops the development of an idea too early.

Anyway – thanks for a really considered analysis. I agree there’s far too little of this, and it makes a refreshing change from the media coverage, which ranges from the quite good to the pretty poor, but is almost always fairly shallow.

My kin in the neuroimaging realm need to be acutely aware of our demographic biases when we’re showing how it’s the XYZ network of colored blobs that indicates you’re better at playing the piano left-handed after you’ve been watching Simpsons videos in the scanner for half an hour. FMRI results in particular tend to get generalized out of all proportion. Caveats abound, not least that one about the test subjects lying in a dark, confined tube with clanking noises all around…

Great reporting on how science works. Can we really have too much of this? I don’t think so. Although I’m curious how the original hypothesis came about.

Seriously though, good catch. Sloppy science, or at least reportage, does no one any good.

Just a tangential point also about the dangers of making apparently definitive conclusions. You write “Psychology researchers know all these arguments, and that’s why they’re cautious about drawing simple conclusions from single studies.”

So why is it that clinical psychologists make snap conclusions from single studies (i.e. patients) all too often and sometimes to the harm of the ‘study subject’?