PSY202H5 Lecture Notes - Lecture 2: Variance, Null Hypothesis, F-Distribution
Document Summary
Like a t-test but more flexible: t-test: difference between 2 means, anova: difference between 2 or more means. All between-subjects anovas (1-way, 2-way, etc. ) have 1 dv, which is quantitative (eg. interval/ratio), and 1 or more iv, which are qualitative (eg. nominal) To test null, we calculate two different estimates of the same thing- population variance. One estimate is independent of whether the null is true or false. Like independent sample t-tests, anova assumes the variance within each group is the same, even if the null is false. One estimate of variance is the mean within-group variance (does not depend on if null is true or false) We look at the ratio of out 2 estimates of variance. A ratio of variances has an f distribution. F distribution differs according to degrees of freedom in the numerator (k-1) and degrees of freedom in the denominator (n-k) If calculated f is bigger than critical f, then reject the null.
