They interviewed 966 people with cancer. Presumably those people already had a tumor in at least one side of their brains.
They also interviewed 1716 people without cancer. Presumably these people didn't have a tumor in their brains.
So they find an increased risc of tumors in the side of the head where people held their mobile, but not, presumably, any correllation with how frequently they used their mobile.
This means either
A: If you use a mobile at all you're at risc. The risc doesn't increase with use.
B: Biased reporting, people claim to hold the mobile to the side of their tumor.
C: Biased reporting, people who use the mobile a lot reports the same amount of use as those who don't
or as you said: The sample is horribly biased from the start.
A better way to do this is finding the average mobile use in the population as a whole, then compare that with the mobile use of victims. Doesn't avoid the bias, though.
Slashdot Mirror
This site claims to have a solution of sorts.
I don't Know. While I will agree that it's [deleted by cencor], Disney has produced some new IP.
Like http://psc.disney.go.com/disneychannel/brandyandmr whiskers/index.html
I mean, like, no-one would pay money for something like that, would they?
The E Ink company sells a prototype http://www.eink.com/kits/index.html for $3,000 running Linux! With Bluetooth, USB, and a MMC card reader.
But you're right, a comsumer product should also be a tablet.
I want one too, but only if it can read HTML.
If it can't access Gutenberg http://www.gutenberg.org/ or the Baen free library (and subscriptions) http://www.baen.com/library/, well, what use is it?
But wouldn't RSS imply the ability to display HTML as well?
They interviewed 966 people with cancer. Presumably those people already had a tumor in at least one side of their brains. They also interviewed 1716 people without cancer. Presumably these people didn't have a tumor in their brains. So they find an increased risc of tumors in the side of the head where people held their mobile, but not, presumably, any correllation with how frequently they used their mobile. This means either A: If you use a mobile at all you're at risc. The risc doesn't increase with use. B: Biased reporting, people claim to hold the mobile to the side of their tumor. C: Biased reporting, people who use the mobile a lot reports the same amount of use as those who don't or as you said: The sample is horribly biased from the start. A better way to do this is finding the average mobile use in the population as a whole, then compare that with the mobile use of victims. Doesn't avoid the bias, though.