I am going to use a repeated measures ANOVA design in my subsequent research (3 x 3), however to test my main hypothesis I need only one planned comparison (contrast). How I can calculate the sample size needed to detect an effect at a given power of 0.80 in that particular planned comparison?
Should I use a simple calculation for un-paired t test?
"Should I use a simple calculation for un-paired t test?"
You should followed a similar way, but using F de Snedecor test, and solve in order to the "n" needed to obtain 0.8.
Have a nice time
Helena
Thanks Helena, thanks Sofia,
I am wondering whether I need sample size calculation based on ANOVA interaction. My goal is to not look at the explorative post-hoc analysis, rather, I plan to perform two contrasts comparisons in order to check my main hypothesis.
Best
Wacław
Dear Kukuczi,
I will look at Gpower as well, and use Lmatrix to do the contrasts.
have a nice time
Helena
Dear Waclaw, sample size determination for one-, two-, and three-way ANOVA where at least on factor is fixed as well as or interactions of fixed effects can be found in our book "Optimal Experimental Design with R" (Rasch, Pilz,Verdooren,Gebhardt), 2011, Taaylor and Francis together with the R-programs.
For one-way ANOVA you can use
To experimental designs one then comes via the command
install.packages(“OPDOE“)
and
library(“OPDOE“)
Now one can calculate the sample size for analysis of variance (or for short Anova) via size.anova and find help by
help(size.anova).
Example 5.4 Determine und for a = 4, α = 0.05 β = 0.1 and τ=δ/σ = 2 !
With OPDOE of R
> size.anova(model="a", a=4, alpha=0.05, beta=0.1, delta=2, case="minimin")
n
5
> size.anova(model="a", a=4, alpha=0.05, beta=0.1, delta=2, case="maximin")
n
9
Thank you very much for all responses. I am going to look at both, G*power and 'R' approach.
Hello, Wacław
This website may be helpful to you:
http://www.sample-size.net/
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size