I've put PEAR lab as a possible topic for mythbusters to tackle on their message boards:
[URL]http://community.discovery.com/eve/f...6/m/2771958968[/URL]
That's great! I don't know if the hosts of Mythbusters normally do this but it would also be great if they could verify on their show that:
- No one has succeeded in replicating what PEAR claimed they could do.
- PEAR never published their procedures.
- PEAR despite "inviting" people to use their lab - never actually gave anyone access to their experiment's results, but would add them in anonymosly to their database. Only the PEAR managers had access to each experiment's result and they would not share this information -- just the results as they had chosen to calculate and as they had chosen to select. (I am assuming that they did have access to each experiment's results, because otherwise how could they have claimed results varied based on gender?)
Also, shortly after you posted at Mythbusters someone posted this:
gatergrater said:[url said:http://www.princeton.edu/~pear][/url]
The observed effects are usually quite small, of the order of a few parts in ten thousand on average, but they compound to highly significant statistical deviations from chance expectations.
In other words, they're taking the tiny deviations from ideal statistics which are always observed in real world experiments, adding them, and calling them significant now that they're bigger. This is fundamentally flawed. Instead, they should re-calculating the statistics using the larger sample size.
I can't help but notice that the site is very long on rhetoric, yet extremely scant on actual data or results.
…
Emphasis mine.
Even based on the tiny bit of knowledge that I have in statistics, I know that this is just plain wrong.
Last edited: