One Way ANOVA: k = number of groups i = group number; j = number of samples in ith group (group size) Yij follows a normal distribution Assumptions: Normal distribution for Yij Distribution of Yij has common variance and equal variance Observations are independent 5 steps: 1. Ho = all mus equal; Ha = not all mus equal 2. State Alpha 3. F Test (F~F I1,nIer the null) 4. Interpret (e.g. Given Fobs > F*, we reject the Ho) 5. Conclusion (e.g. That is, there is strong evidence to suggest Group A > Group B) Pairwise Test: Comparing two groups within an ANOVA H0: mu2=mu3 vs HA: mu2 == mu3 If using CI, reject Ho if CI does not cover 0 CI= Calculate variance, standard deviation using all groups (group pooled sd or variance) (Continue until all groups are covered, not just the compared groups, and subtract by of group) Df still = n1 Linear Combination Gamma: SE(Gamma): H0: gamma = 0 vs Ha: gamma == 0 Number of pairs for comparison: Where I = number of groups Standard Deviation: