One of the first steps in planning a new study is determining an appropriate sample size. A sample size should be large enough to have a high probability of detecting an effect of a treatment but small enough that it is within the confines of the study?s budget and minimizes the potential risks to human and animal subjects. Within this workshop we will investigate the relationship among sample size, effect size, type I error rate, type II error rate, and power. Examples will be worked out for hypotheses that can be answered using t-tests, ANOVA, and regression using the G*Power software. How to calculate and implement a design effect to account for a multilevel design will also be covered in the workshop.
Relevant Software: