Psychology 3780F/G Lecture Notes - Lecture 8: Sample Size Determination, Type I And Type Ii Errors
Document Summary
Convenience sampling taking what you can get. Purposive sampling non-representative subset of some larger population, constructed to serve very specific need or purpose. Sample size: p-values are impacted by sample size, effect sizes are not impacted by sample size. Likelihood of deciding that there is an effect when effect is present. Larger the effect size predicted, the fewer # of participants needed: much harder to get significant results w/ small # of participants. Using more potent interventions that have bigger effects. Reduce measurement error (use highly valid outcome measures) Raise alpha level (but only if making type i error is highly unlikely) Aim for beta . 2, indicating . 80 power. When studies are underpowered (not enough participants), run risk of missing significant finding (type i error)