2·
2 months agoNo, it’s not. Standard error of the mean estimate is SD/sqrt(n-1), so it depends on the signal/effect to noise ratio how much you can learn from it. SE of the mean difference for a between subjects design is slightly different but not miles off from that. You can squeeze even more info out of it by using within subjects designs.
What makes you think a sample size of 50 is irrelevant?
Correct, some contracts cannot be done. Depending on the specifics, the illegal parts of the contract are unenforceable, or the entire thing is null and void. Employers often hate this, because it limits the exploitation they can enact upon employees.