PSYCH 100A Lecture Notes - Lecture 12: Mean Absolute Difference, Null Hypothesis, Analysis Of Variance
Document Summary
You need to make sure that you average out the standard deviation to get the mean square. Due to random chance alone, drawing three samples of n=5 from a population will produce mean differences (sampling error) Anova uses sums of squares and variances to quantify mean differences and chance differences. Divide the total variability in the dependent variable into variation due to the independent variable (mean differences) and left over (residual) variability. Between-group variability is variation due to the independent variable (mean differences) Mean difference variation is computed by comparing each treatment group mean to the overall mean. You compare the group mean to grand mean. If it was a large effect, you would expect a big difference between the grand mean and the group means. Within-group variability is leftover (residual) variation that is not due to the independent variable. Captures the leftover variation not due to the independent variable. Anova calls the leftover (residual) variation error.
