3 ms·

Placebo effects in cognitive training [pdf]

39 points by lainon 10y ago

robert_tweed 10y ago
At first glance this seems like a horribly flawed study. It seems misleading to call this a placebo effect when it is clearly the result of selection bias. The paper makes it clear that this is intentional, but I'm not sure they've really shown anything novel here, just demonstrated what should already be well known. I.e., badly designed studies may contain bias.
disgruntledphd2 10y ago
This is actually not a bad study. First, the bad stuff: - Only 50 participants (they waffle about power, but its unconvincing) - Over analysis of such a small dataset (they report testing lots of issues, and they presumably did a bunch more tests). On the good side: - They showed a 1pt improvement on a good measure of intelligence (raven's matrices), as a result of a differing recruitment strategy (scroll through the paper, there's a half-page image showing these). - All of the experimenters were blind to both study purpose and condition (which is sadly rare in psychology). - This is a really interesting effect. It would be one thing if it were a social measure, but raven's matrices are an IQ test. This would seem to support the notions around malleable intelligence (Carol Dweck et al), and is an interesting experimental effect. That being said, the sample size was really small, and I probably won't believe this until I see it replicated a few times (though I've seen similar placebo effects on IQ tests before in past research, which I always thought sounded more like stereotype threat).
SubiculumCode 10y ago
I abandoned my cognitive training research plans after some piloting of a task. I had a number of data driven reasons for it, but a personal one was that I doubted that cognitive training would ever be more effective for the majority of people than spending that time exercising, and following better sleep habits. Sure one can do all three, but for many there are real time constraints (anyone with kids here agree?) and choices have to be made...and choosing to sit down in front of a computer for cognitive training is probably not the right choice.
- SubiculumCode 10y ago
  Actually given the relatively ample evidence of the cognitive boost associated with exercise, an exercise condition would make an excellent active baseline to test cog training against. It would make a great study.

39 points by lainon 10y ago

At first glance this seems like a horribly flawed study. It seems misleading to call this a placebo effect when it is clearly the result of selection bias. The paper makes it clear that this is intentional, but I'm not sure they've really shown anything novel here, just demonstrated what should already be well known. I.e., badly designed studies may contain bias.

disgruntledphd2 10y ago

This is actually not a bad study. First, the bad stuff: - Only 50 participants (they waffle about power, but its unconvincing) - Over analysis of such a small dataset (they report testing lots of issues, and they presumably did a bunch more tests). On the good side: - They showed a 1pt improvement on a good measure of intelligence (raven's matrices), as a result of a differing recruitment strategy (scroll through the paper, there's a half-page image showing these). - All of the experimenters were blind to both study purpose and condition (which is sadly rare in psychology). - This is a really interesting effect. It would be one thing if it were a social measure, but raven's matrices are an IQ test. This would seem to support the notions around malleable intelligence (Carol Dweck et al), and is an interesting experimental effect. That being said, the sample size was really small, and I probably won't believe this until I see it replicated a few times (though I've seen similar placebo effects on IQ tests before in past research, which I always thought sounded more like stereotype threat).

SubiculumCode 10y ago

I abandoned my cognitive training research plans after some piloting of a task. I had a number of data driven reasons for it, but a personal one was that I doubted that cognitive training would ever be more effective for the majority of people than spending that time exercising, and following better sleep habits. Sure one can do all three, but for many there are real time constraints (anyone with kids here agree?) and choices have to be made...and choosing to sit down in front of a computer for cognitive training is probably not the right choice.

SubiculumCode 10y ago
Actually given the relatively ample evidence of the cognitive boost associated with exercise, an exercise condition would make an excellent active baseline to test cog training against. It would make a great study.