Steven Samuels wrote:
>David is partly right.
I think you meant me.
>Welch's test uses separate standard deviations to compute the
denominator of the t-statistic, and, as David says, this is the same
denominator as that computed by the robust option in -reg-. However -
reg- with the robust option uses the usual d.f. for the test (n-2),
but Welch used a corrected d.f.>
Actually you are given three options. One is the default df adjustment. The others use the hat matrix in various ways.
Hmm, I wonder if the other adjustments end up being general versions of common adjustments used in ANOVA Land by ANOVA Man, e.g., Huynh-Feldt for RM? If so this would be a very nice simplification for teaching, replacing a laundry list of adjustments that appear ad hoc from the student's perspective with a lecture on the effect of heteroscedasticity and a useful remedy for it. Time to stare at the equations....
JV
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
