PSYC 3130 Lecture Notes - Lecture 10: Sphericity, Statistical Hypothesis Testing, Statistical Power

• School: University of Windsor
• Department: Psychology
• Course: PSYC 3130
• Professor: Ken Cramer

Document Summary

Variables do not need to be independent. There is no limit on the number of measures (or rather times we measure participants) Under null, assumption that there is no change anywhere in the system. Alternatively, there is a difference somewhere in the system. Greatest source of error is individual differences. F gets larger when treatment effects are large and error is small increase in statistical power (greater f statistic) There are no groups of different people in ranova. Must name a factor and select number of levels, then define it. Choose estimates of effect size in spss when performing ranova. Significance in within-subject contrast will indicate if relationship is more likely linear or quadratic (closer to . 05)