I am trying to conduct power analysis in Stata using the stored results from the ttest command.
What I did exactly is as follows:
. ttest r, by(predictor)
Two-sample t test with equal variances
------------------------------------------------------------------------------
Group | Obs Mean Std. Err. Std. Dev. [95% Conf. Interval]
---------+--------------------------------------------------------------------
0 | 8 .49 .0368394 .1041976 .4028886 .5771114
1 | 8 .595 .0547396 .1548271 .4655613 .7244387
---------+--------------------------------------------------------------------
combined | 16 .5425 .034635 .13854 .4686772 .6163228
---------+--------------------------------------------------------------------
diff | -.105 .0659816 -.2465165 .0365165
------------------------------------------------------------------------------
diff = mean(0) - mean(1) t = -1.5914
Ho: diff = 0 degrees of freedom = 14
Ha: diff < 0 Ha: diff != 0 Ha: diff > 0
Pr(T < t) = 0.0669 Pr(|T| > |t|) = 0.1339 Pr(T > t) = 0.9331
. display r(mu_1),r(mu_2),r(sd_1),r(sd_2),r(p_l)
.49 .595 .10419762 .1548271 .06692561
However, when I run the following command:
power twomeans r(mu_1) r(mu_2), sd1(r(sd_1)) sd2(r(sd_2))
I always receive the error below:
means must contain numbers
I thought the stored results r(mu_1) etc. are numbers?
Using the auto dataset as a toy example:
sysuse auto, clear
ttest mpg, by(foreign)
Two-sample t test with equal variances
------------------------------------------------------------------------------
Group | Obs Mean Std. Err. Std. Dev. [95% Conf. Interval]
---------+--------------------------------------------------------------------
Domestic | 52 19.82692 .657777 4.743297 18.50638 21.14747
Foreign | 22 24.77273 1.40951 6.611187 21.84149 27.70396
---------+--------------------------------------------------------------------
combined | 74 21.2973 .6725511 5.785503 19.9569 22.63769
---------+--------------------------------------------------------------------
diff | -4.945804 1.362162 -7.661225 -2.230384
------------------------------------------------------------------------------
diff = mean(Domestic) - mean(Foreign) t = -3.6308
Ho: diff = 0 degrees of freedom = 72
Ha: diff < 0 Ha: diff != 0 Ha: diff > 0
Pr(T < t) = 0.0003 Pr(|T| > |t|) = 0.0005 Pr(T > t) = 0.9997
You need to explicitly tell to power that the returned items from ttestare local macros:
power twomeans `r(mu_1)' `r(mu_2)', sd1(`r(sd_1)') sd2(`r(sd_2)')
Performing iteration ...
Estimated sample sizes for a two-sample means test
Satterthwaite's t test assuming unequal variances
Ho: m2 = m1 versus Ha: m2 != m1
Study parameters:
alpha = 0.0500
power = 0.8000
delta = 4.9458
m1 = 19.8269
m2 = 24.7727
sd1 = 4.7433
sd2 = 6.6112
Estimated sample sizes:
N = 46
N per group = 23