Topic:
The same four proportional weighting factors work for each 00/01/10/11 when 0.25 is subtracted from each !!!
Replies:
506
Last Post:
Nov 20, 2012 9:21 PM
On May 15, 7:28 pm, djh <halitsk...@att.net> wrote: > Thank you very much for explaining that the logical inputs relate > the way they do because the data were derived from a double > "median" split on the residuals of the "u" and "e" correlations. > > And thank you very much also for computing and posting the co- > variances. > > I'm really glad to have a "record" of the co-variances for the > adjusted input because as I mentioned to you, one of the new data > passes I'm doing now is on the ORIGINAL study and control group > but with the new constant-ratio length intervals (and some minor > inaccuracies removed).
The logs of the geometric means of the new constant-ratio length intervals are close enough to being equally spaced (corr = .9992, sqrt[1-corr^2] = .04) that it should be OK to use the integers 1,...,12 (or any other convenient set of 12 equally-spaced values) as predictors instead of the actual lnGM's. That should simplify the clerical work.
> > And of course, it will be instructive to compare co-variances from > this new pass to the co-variances you've posted above, particular > because w is such an important predictor.
