Psycho-Babble Medication | about biological treatments | Framed
This thread | Show all | Post follow-up | Start new thread | List of forums | Search | FAQ

statistical inference

Posted by llrrrpp on August 24, 2006, at 17:11:40

In reply to Re: Interesting Article on Dopamine! xbunny, posted by Phillipa on August 23, 2006, at 19:49:05

Thanks for the article, Tom.

Chris Frith is doing a pretty poorly designed study, if we are to take this NewScientist article as a good representation of his actual experiment.

Let's assume that his study does not have any major confounds (design problems) Ideally we'd like to have a sense that the behavior in question (gambling addiction) is affected in a dose-dependent fashion to the pdockery treatments. For instance, does the same person react differently to a low dose and a high dose? Are these effects in the predicted direction? This might be an important follow-up study.

Now let's talk about the possible (and likely) problems with this study. A VERY important control in this type of work is using a secondary task, for instance, a task that requires decision making in the absence of learning. Are these results in the article merely an effect of slower reaction time and diminished accuracy. Would all kinds of tasks show these effects on haldol? on L-Dopa? If so, these results are not unique or special to the effects of dopamine on gambling and learning behaviors.

Also, what are the types of errors? are they errors where participants are more cautious? or are they errors where participants are more generous and inclusive? might one drug cause one type of error, and another type of drug cause the other type of error? (regardless of the type of cognitive task being administered?). If so, it must be demonstrated that a particular type of error bias cannot account for the "gambling addiction" results.

Anyways, these are VERY basic experimental design issues. Any cognitive psychologist will learn about these types of concerns by the time they finish their masters degree (or even sooner).

Sadly, Frith's study is similar to almost ANY study I've ever read that examines cognitive effects of psychopharmacology. It kills me, every time, to see what is potentially a very clean elegant study completely botched, because the people running the study negelected basic experimental design principles and logic. I'm pretty shocked to see this stuff from this lab, however. Frith is a researcher with lots of resources and money, and I would have expected him to use these with more discretion.

Everyone seems to be jumping on the neuroeconomics bandwagon nowadays. It's the hot new field, and seems to attract people who want to make a bang, without having to worry about following basic principles of statistical inference...

Hey, you? Wanna get published? take a group of 10 college kids. give them some marijuana, and then have them play a violent videogame. then ask them if they wanna buy it. Now publish a study that says that marijuana increases violent videogame purchasing. (oh, small problem-- you have to get the IRB (university/hospital ethics review) to approve your study, and you have to get informed consent from the college kids.) sign me up!

my rant. Can you tell I got my provigil in the mail today?? HALLELUJAH!!! my cognitive faculties have returned!





Post a new follow-up

Your message only Include above post


Notify the administrators

They will then review this post with the posting guidelines in mind.

To contact them about something other than this post, please use this form instead.


Start a new thread

Google www
Search options and examples
[amazon] for

This thread | Show all | Post follow-up | Start new thread | FAQ
Psycho-Babble Medication | Framed

poster:llrrrpp thread:679343