Any suggestion on calculating sample size for a superior RCT in R or SAS?
I have used the 'pwr' package which gives me a different result from the one I got from an online calculator (Sample size calculator (riskcalc.org)).
my parameters are
control- change in mean- 20+/- 5
treatment- change in mean- 15 +/- 5
drop out -20%
power-0.9
Any suggestion would be appreciable.
Hello Hanish,
Presuming that your threshold for clinical significance to detect between groups is 5 units or more (20 vs. 15), and that the SD for each group is 5 points, then your study is intended to determine if a group difference of at least 1 SD (5 units or more on a scale having an SD of 5 units) exists.
To mimic the two-stage process outlined by Chou & Liu (2004), then you could estimate N by evaluating a directional hypothesis test at target risk level of alpha/2 (see Article Comparisons of Superiority, Non-inferiority, and Equivalence Trials
)Via the freely available program, g*power (t-test, difference between two means), ES = 1.0, alpha = .025, power = .90, allocation ratio (n2/n1) = 1, the resultant N is 46 cases. With maximum attrition of 20%, that implies a starting N of at least 58 cases (29 per batch).
Good luck with your work.
Certainly! Calculating sample size for a randomized controlled trial (RCT) involves considering several factors, such as the effect size, variability, significance level, power, and dropout rate.
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Modeling
- Analyses of Variance
- ANOVA