Right. I will just take the last because its the most fun. PSON, Can you explain why he use n=68 in this line?
- Z = 68 1/2 * (.291) = 2.40
(but thats just nit-picking)
The most significant problem is that he regress a complicated measure from a (self-invented) model against real data. Which mean his results are only valid if the parameters in the model is obtained WITHOUT looking at the final result. Ofcourse we never know where he gets the parameter values from, nor where the model come from, so i fear the worst.
Secondly he should ofcourse have regressed the other observations (independently) against the final result without using the model, just to make sure there are no trivial corrolations in his dataset.
for those two reasons alone he would never pass peer review.
So what are your thoughts on that, PSON?
the above is an answer to your post 1597. I dont have time to review the other right now.
I can ask in another way. why does he use this model, and where does he get the parameter values from?
I = c * [(log(SS+1) / 2.5) 4 * (log(RH mean +1) / 2) 2 * ((deltaRH + deltaRH max ) / (2 * deltaRH max )) * cos(%CC * (pi / 200))] + 10