PSY 303 Lecture Notes - Lecture 15: Pairwise Comparison, Studentized Range, Analysis Of Variance
Document Summary
Anova and hypothesis testing: problem set due next friday! (at least! Estimates of the magnitude of the difference between conditions. Different from statistical significance which is influenced by sample size. A very small effect can be statistically significant with a very large sample size. this presents a problem. So its important to report effects size along with sample size to prevent error or skewing of data. Statistical significance- effect size x sample size: effect size equation. Eta squared n^2 proportion of total variance in the dv explained by the iv. Exactly same as the r^2 concept in correlation/ regression. Helps get away from arbitrary value of . 05. Small= . 01, medium=. 06, large=. 14-----won"t be tested on. Put this in last column of source table==eta^2. Can therefore say there is a statistically significant effect that is what p value tells. If eta=. 33 ==33% of data is because of our manipulation us.
