Marco’s blog

Journals’ impact factors are not as important as they used to be. Nevertheless I found an intersting parer on publication bias by Marcus Mufano (with a stress on the last o I think). He has lots of interesting papers, here is his homepage:  http://psychology.psy.bris.ac.uk/people/marcusmunafo.htm.

Here is a quick look at just one graph. A genetic link has been established so we can tell when a paper is reporting data that are biased. The intersting thing is that the bias gets stronger the higher the impact factor of the journal. Possibly people are trying too hard to sell their findings and make them appear exciting.

The government plans to go ahead with the Heathrow airport expansion and with airport expansion across the country even though this means we’ll have no hope of meeting our climate emission targets.
I’ve joined the airplot:

APS President Walter Mischel in the APS Observer has a nice discussion of the “toothbrush problem” in psychology, (it’s unclear who was the person who coined the name). Basically no self-respecting scientist would use someone else’s model. This leads to lots of pet models and little attempt to integrate knowledge. In his conclusions Mischel says “for psychological research to flourish and develop into an increasingly cumulative basic science, there are some fundamental requirements. It’s essential to develop and use common shared tools and a common language, so that replication, and building on solid work becomes accepted practice and is valued.” It is also essential, he says, to publish failures to replicate. Wise words but hard to implement, given the usefulness of having one’s own toothbrush in terms of career. but also given the way that journals work. Nevertheless it is nice to see the issue given such a high profile discussion.
MB

Constantstimuli is the name of a C++ generic program to run visual perception experiments. Martin and I have finished putting it together recently and I use it in the lab. Anybody interested in the code can write to me.

There is something about programming that I always found attractive and rewarding. For someone with a short attention span, debugging (which for the non-pro like me is most of the programming time) can be done by a series of quick series of hypotheses and tests to see what goes wrong where, so that there is a continuous interaction with the machine. I like that feeling of crafting something out by small steps, while keeping the big plan in your head, which is what a sculptor presumably must do.

Anyway, I am very happy with Constantstimuli and perhaps it will get even better in the future. We also wrote a related program called Adaptivestimuli for staircases (when a stimulus is changed on the basis of the observer’s response). It works but may need more testing and fine tuning.

cheers, MB

A nice summary on the BPS blog of one of our research papers:
http://bps-research-digest.blogspot.com/2008/08/youre-not-as-big-headed-as-you-think.html

Update: I had put the wrong link! Fixed now