PSY 350 Lecture Notes - Lecture 27: Central Limit Theorem, Standard Deviation, Statistical Inference
Document Summary
If we have an estimate of the effect size (or possible effect sizes), we can use power analysis to figure out how many participants we need to get an acceptable level of power. To find that a correlation of . 30 is significant at = . 05, i need at least 85 participants in my sample. To find a significant difference between two groups at = . 05, where i expect the difference to be at least 0. 5 sd, i need at least 63 participants per group. For simple analyses, there are straightforward functions to calculate these things. Grad schools and good journals expect you will do a power analysis before collecting data. Review the general framework for any kind of inference (hypothesis) test. And know what to do if they"re not met! Understand how the components of the z-score affect the likelihood of rejecting the null. Differentiate one- and two-tailed hypothesis tests and when to use each one.
