Presented November 13, 2015
Effect Size & Power Analysis + G*Power
What is Effect Size?
- A quantitative (numerical) measure of the strength or magnitude of a relationship
- The criterion for evaluating the strength of a statistical claim
- A larger absolute value indicates a stronger effect
- Plays an important role in power analysis, sample size planning, and meta-analysis
- Examples: correlation between 2 variables, difference in group means, risk ratio, ect.
What is Power Analysis?
- Answers the question: how large does my sample need to be?
- Should always be done while designing an experiment or survey
- Four pieces: sample size, effect size, significance level, & Power (given any 3, you can solve for the 4th)
What you will be able to do after this workshop
- Know how to determine effect size for various types of study designs
- Convert between various measures of effect size
- Understand the relationship between the four components of a power analysis
- Use G*Power to determine required sample size ‘a priori’ when planning a study
- Use G*Power to determine observed power after a study is complete
How to obtain G*Power
Directly available for FREE for PC or MAC at http://www.gpower.hhu.de/