Ok, I’m still not sure, but I think we’re warm.
Hypothetically, if rasch evaluated to 0 or negative, I suppose I’d get a log(0) = -inf
or more likely… where I transpose my original data containing NaNs and multiply by the log of rasch
Ok, I’m still not sure, but I think we’re warm.
Hypothetically, if rasch evaluated to 0 or negative, I suppose I’d get a log(0) = -inf
or more likely… where I transpose my original data containing NaNs and multiply by the log of rasch