|
Size: 526
Comment:
|
← Revision 4 as of 2013-03-08 10:18:09 ⇥
Size: 597
Comment: converted to 1.6 markup
|
| Deletions are marked like this. | Additions are marked like this. |
| Line 9: | Line 9: |
| alpha <- 0.05 | beta <- 0.05 |
| Line 20: | Line 20: |
| (pow - (1 - pt(abs(qt(alph/2,(2*n)-2)),mdiff/sqrt((2*sd*sd)/n*(1-rho))) )^2 | (pow - (1 - pt(abs(qt(beta/2,(2*n)-2)),2*n-2, mdiff/sqrt((2*sd*sd)/n*(1-rho)))) )^2 |
| Line 26: | Line 27: |
| pow <- 0.8 alpha <- 0.05 mdiff <- 2 sd <- 4 rho <- 0.5 |
|
| Line 27: | Line 33: |
Sample sizes for paired t in R
mdiff is the difference between the two variables with correlation, rho, type I error, alpha, and (common) sd.
[COPY AND PASTE INTO R AND ADJUST INPUTS AS DESIRED]
pow <- 0.8 beta <- 0.05 mdiff <- 2 sd <- 4 rho <- 0.5
[THEN COPY AND PASTE THE BELOW INTO R] [OUTPUTS NUMBER OF PAIRS]
fn <- function(n) {
(pow - (1 - pt(abs(qt(beta/2,(2*n)-2)),2*n-2,
mdiff/sqrt((2*sd*sd)/n*(1-rho)))) )^2
}
nout<- nlm(fn,2)
nout<- trunc(nout$estimate+1)
print(nout)
pow <- 0.8
alpha <- 0.05
mdiff <- 2
sd <- 4
rho <- 0.5