PSYC 3000 Lecture Notes - Lecture 5: Effect Size, Mean Absolute Difference, Sampling Distribution
Document Summary
Z-scores are getting higher when moving away from the middle: high numbers mean very unlikely to happen -> low probability, 1 larger portion = smaller portion. Clt if your sample size is 30 or more then you are allowed to assume your sampling distribution (unicorn) is normal: even if population is wonky, even if sample data is wonky. If you know your population is approximately normally distributed ?: assume sampling data is also normal. If sample data is approximately normal (this we can test we have data) then we are allowed to assume the population is normally distributed (which lets us assume a normal unicorn) If none of the above, then we need non-parametric test (not borrowing information from the normal curve) Comparing a sample mean (continuous variable) to a hypothesized population mean (guess: default is spss is a zero (is this logical, may or may not be 0.