Supplemental Results
(Posted 12/18/03 20:15 PST)
I. MAIN CASES: Files 1, 2, and 3 contain both (a) equal and (b) unequal group sample sizes
for the normal, gamma, lognormal model, and dependent data. All files are in PDF format.
1. P-values based on t-statistics with pooled variance
[Estimate of Pi0]
[Power]
2. P-values based on permutation
[Estimate of Pi0]
[Power]
3. P-values based on a Welch correction to DF for unequal variance
[Estimate of Pi0]
[Power]
II. SPECIAL/ADDITIONAL CASES: (a) equal and (b) unequal sample sizes
4. Contant variance, fixed alternative mean: case (A)
(a) [Estimate of Pi0]
[Power]
(b) [Estimate of Pi0]
[Power]
5. Contant variance, mixed alternative means: case (B)
(a) [Estimate of Pi0]
[Power]
(b) [Estimate of Pi0]
[Power]
6. Smaller effect size: case (A)
(a) [Estimate of Pi0]
[Power]
(b) [Estimate of Pi0]
[Power]
7. Larger effect size: case (B)
(a) [Estimate of Pi0]
[Power]
(b) [Estimate of Pi0]
[Power]
8. Mixed null variances
(a) [Estimate of Pi0]
[Power]
(b) [Estimate of Pi0]
[Power]
9. Main case, m = 3000
(a) [Estimate of Pi0]
[Power]
(b) [Estimate of Pi0]
[Power]
|
