Skepticism, Science & Pseudoscience

These guys are exactly like Seneff, with her over wrought claims Roundup causes Parkinson's Disease and that half our children will be autistic by 2050.

It would be useful for them to run an actual analysis on their data and estimate how many spurious correlations they are likely to pull out of the muck under the conditions they use.

In Seralini's original paper, he measures a metric feltchton of parameters and fishes out 40 that he claims are significant at the (p < 0.05) level. Setting aside the fact that this criterion is known to be pretty sloppy, that leaves him with an expectation of 25 +/- 7 spurious results. He builds his entire pyramid of global death on a foundation of about 15 actual results (and plausibly as few as 8) in a context where he doesn't know which 15 of those 40 numbers are real and which 25 are noise and mean nothing.