SOC202H1 Lecture Notes - Lecture 7: Sampling Distribution, Interval Ratio, F-Distribution
Document Summary
What if we have more than 2 groups. Ma(cid:374)(cid:455) ti(cid:373)es (cid:449)e"d like to (cid:272)o(cid:373)pare (cid:373)ore tha(cid:374) 2 groups. We may be tempted to consider conducting multiple t-tests. However, carrying out multiple tests comparing each of the means would present a big problem. If alpha was set to 0. 05 for each of these 10 hypothesis tests, we would have a 40% chance of making a type 1 error considering a difference statistically significant when it simply arose by chance. The analysis of variance procedure overcomes this problem a simple way of comparing multiple means. You can think of anova as extension of t-test for more than two groups. Anova asks (cid:862)are the differences between the samples large enough to reject the null hypothesis and justify the conclusion that the populations represented by the samples are different? (cid:863) The h0 is that the population means are the same: H0: 1= 2= 3 = = k.
