hmm. isn't 1 -1 the correct contrast for patient vs. controls? it woudl seem that 1 0 woudl assign 0
to the second group and thus test whether the patients were greater than zero. wouldn't 1 -1 test
for patients greater than controls and create a vector that summed to zero?
as far as t and F go, as i understand it, with two groups they are equivalent in the sense that a the
resulting t value would be the square root of F. in terms of why it doesn't work in this case, the error
message woudl be informative.
|