
125Exploratory Analysis and Introduction to Inferential Statistics
If we relax the effect size to the difference 45.7 − 44, i.e., a delta ∼2, the power goes up to 0.79.
We conclude that for this sample the t-test would have relatively low power given α = 0.01.
We can now determine how many more observations would be required to raise the power to 0.9
assuming the same variability, same α, and same effect size of delta = 2.
>power.t.test(power=0.9, delta = 2, sd=sd(poro$V1), type=”one.sample”,
alternative = “one.sided”, sig.level=0.01)
We obtain
One-sample t test power calculation
n = 12.30839
delta = 2
sd = 1.702939
sig.level = 0.01
power = 0.9 ...