[Synoptic-L] Count correction and upload
- The counts I gave before were for E-I only. Sorry about that. The correct
counts for A-I are below:
T222 T211 T112 T212 T221 T122 T121 T022 T012 T021
820 653 891 107 547 316 733 146 190 226
T220 T120 T210 T020 T202 T201 T102 T200 T002
580 574 403 356 801 684 618 1908 2927
I also uploaded two files:
One is just raw (frequency - means) for the Greek words that are being used.
The other is a run with the zeros replaced with "missing values".
I had to go back to subtracting the total frequency (as I was doing in
Excel), instead of partial correlations because SAS insisted every variable
should have a non-missing value for the observation to be used in a partial
correlation. I don't see why that is technically necessary. I think only the
partial variable should be required to be non-missing. It probably has to
with the fact that it really does a regression in the process of computing
the results. In any case, this left only 17 words, so I used a different
The results are similar to what they were. 212 significance is lower, as
The results now list 3 numbers. The first is the correlation coefficient.
The second is the probability that it is really zero (or a value on the
other side of zero). The third is the number of Greek words that were
non-zero in both categories, thus contributing an observation to the
Ph.D. Management Science candidate
Synoptic-L Homepage: http://www.bham.ac.uk/theology/synoptic-l
List Owner: Synoptic-L-Owner@...