Comment: Re:And it took this long to "make the connection"? (Score 2) 248
If that's the case then I'd like to see their sample size for full mouth x-rays. I'd bet it's smaller than the sample size for the partial/single bitwing and that could be the source of the vanishing connection. Basic statistics, if a known correlation exists, and you didn't find it in your particular sample check your sample size first. Throwing out the correlation because the smaller sample doesn't show what the larger sample does sounds a bit foolish, and if done that way would result in no correlation ever being found in anything.
Or the sample size for full mouth was larger and a better representation of the norm. Or both were the same just too small to draw a conclusion from.
Either way Lurie is correct in asserting that this type of inconsistency results in an inconclusive study.