Sample size calculation

Confidence intervals and standard error rates have no meaning when the sampling (or at least the assignment to test groups) is not random regarding the factors that you want to test.

Reference: https://www.researchgate.net/post/Do_you_know_any_sample_size_calculation_for_nonprobability_purposive_sampling_technique

It is very useful information! thank you!

There is no way to estimate sample size when the sample is from non-random sampling. One can only assume the use of random sampling to set the minimum. Typically for any evaluation, the assumption of random sampling suggests a size of 780 cases for an expected weak effect size, 123 for an expected medium effect size, and 46 cases for an expected strong effect size, to achieve 95% in confidence and 80% in statistical power. It thus depends on the weakest effect size expected.