I saw on your response to an anon question on the first of march that you are "uncomfortable with a large part of behavioural science". I was wondering if you could elaborate? I'm currently studying behavioral science, and I'm curious what some of the weaknesses others see are. I see some methodological issues with the field, but think it's still quite valuable. I'd like to know where the problems are, so I can make an informed decision about where to put my further study energy.
A few big problems for me with behavioural science are how frequently experiments are poorly designed, the level of control is impossible to assess, and behaviours are subjectively interpreted. I also think that a lot of modelling that takes place in behavioural science tells us very little about the animals they are looking at except that they don't necessarily behave like a computer simulation. Quantifying behaviour in a meaningful, informative, and interesting way is extremely difficult. There is also a problem of perceived significance (and the stupidity of arbitrary significance) and pathetic sample sizes in a LOT of behavioural research. I have read several behavioural studies where the effective sample size was about 3 individuals, because of the number that had to be discarded.
I am not saying it's not a valuable field. I would be the last person to say that a field is valueless, considering how niche my own field is. What I am saying is that in my experience, from the behavioural research I have done, the literature I have read, and the lectures I have sat through, a lot of the conclusions that are being drawn in behavioural fields are stronger than the data warrants. This is something that should be avoided whenever possible in science.
Behavioural science is actually a really exciting field in terms of technological and methodological advancement. Hopefully we will be able to transfer experiments that have previously been done under lab conditions to the field, where control is more difficult but conclusions are more relevant to real behaviours. My partner's research for example has just demonstrated that the gluing of PIT tags to field crickets doesn't affect their longevity in the field, meaning that research on these insects can now be done in a field environment without worrying about the study subjects having lower survival than their unstudied counterparts.
I just urge anyone in behavioural science to go forth with careful consideration of their goals and the ways in which to achieve them in a statistically sound and reasonable way. As I have discovered myself, it is so easy to make a study meaningless by less than careful research design. Avoid that at all costs.
