Python vs G*Power: Sample size calculation for Pearson correlation
Here I compare the power analysis for sample size calculation for Pearson correlation.
The codes can be seen as follows:
import pingouin as pg
Parameters
r = 0.3 # expected correlation coefficient
alpha = 0.05 # significance level
power = 0.80 # desired power
Calculate sample size
sample_size = pg.power_corr(r=r, alpha=alpha, power=power)
print(f"Required sample size: {sample_size:.2f}")
for one-sided test
Define the parameters
r = 0.3 # expected correlation coefficient
alpha = 0.05 # significance level
power = 0.80 # desired power
alternative = 'greater' # alternative hypothesis for one-sided test
Calculate the required sample size
sample_size = pg.power_corr(r=r, alpha=alpha, power=power, alternative=alternative)
print(f"Required sample size: {sample_size:.2f}")
More info about Pingouin
https://pingouin-stats.org/build/html...