It has been demonstrated that predators can select against high growth rates and risk-taking behaviour by eating out the bold, feeding-machine type of fish (Biro et al. 2004; 2006), and the case is the same if the predator happened to be human (Biro & Post 2008).
In a recent paper published in TREE, Biro & Dingemanse (2009) have compiled evidence for animal personality based sampling bias in ecological studies. The story is fairly simple: in many cases, especially when passive methods (like non-baited traps) are used, researchers will end up with far more bold individuals than they would find in a truly representative sample. There are examples listed from sunfish through squirrels to bighorn sheep. A Hungarian colleague of mine is working on similar issues with birds, and the results likely to be just similar.
I guess I don’t have to go into detailed speculations about how many of the published studies can be biased, and how many traits (obviously not only behavioural ones, but almost anything else) could be falsely estimated for the studied populations. Pretty worrying!
/I found no studies about the effect of the sampling researchers’ personality on the quality of samples yet./
Biro PA, Abrahams MV, Post JR, Parkinson EA (2004) Predators select against high growth rates and risk-taking behaviour in domestic trout populations. Proceedings of the Royal Society London B 271:2233-2237. doi:10.1098/rspb.2004.2861
Biro PA, Abrahams MV, Post JR, Parkinson EA (2006) Behavioural trade-offs between growth and mortality explain evolution of submaximal growth rates. Journal of Animal Ecology 75: 1165-1171. doi: 10.1111/j.1365-2656.2006.01137.x
Biro PA, Post JR (2008) Rapid depletion of genotypes with fast growth and bold personality traits from harvested fish populations. Proceedings of the National Academy of Sciences USA 105: 2919-2922. doi:10.1073/pnas.0708159105
Biro PA, Dingemanse NJ (2009) Sampling bias resulting from animal personality. Trends in Ecology and Evolution 24: 66-67. doi:10.1016/j.tree.2008.11.001
4 Comments
Interesting! Have to read it but the basic idea has been around for a good while: in CMR-study literature there is always this discussion about assumptions and ‘trap-happy’ and ‘trap-shy’ individuals.
Anyone who has trapped +few hundred pairs of tits (or flycatchers – if you prefer) knows that the game can be quite different depending on individuals!
Yes, it is pretty logical and one would assume it to happen.
With fish, one might use some hardcore methods (huge seine nets?) where the sampling of a certain water volume can be close to 100%, but I really wonder what kind of method can be used with birds, mammals, or anything terrestrial and larger than small non-flying insects (those can be collected nicely with those vacuum-cleaner type of things)?
If this is to be considered seriously, the taxonomic range of realistic model species used in evolutionary ecology (especially when based on wild populations) can become pretty narrow. But sticklebacks can remain in!
I don’t think that would work with fish either. Different personalities can be distributed spatially (or vertically)differently. In other words, those individuals present in places where huge seine nets could operate may not reflect phenotypic distribution in the entire population.
Do not forget that in some systems more or less the whole population can be sampled. Take for instance the passerine long-term studies. You keep on chasing the difficult ones until you get them. Sure, there is always an odd male that is so superstitious that you cannot fool him, but even they make mistakes once and while.
I do agree that this is an important consideration, but I would not think it would be a serious problem in all instances as a default. But perhaps that is exactly the null-hypothesis once should somehow test/evaluate before jumping to do any other studies in a given system.