MGCR 271 Chapter Notes - Chapter 14: Null Hypothesis, Multiple Comparisons Problem, Analysis Of Variance
Document Summary
Advanced statistical inference often concerns relationships among several parameters. The anova test tests for the equality of the means of any number of populations. It allows any relationship other than (cid:862)all e(cid:395)ual. (cid:863) The anova test is an overall test that tells us whether the data give good reason to reject the hypothesis that all the population means are equal. Contrasts are used to answer speci c questions about group means that can be formulated before the data are collected. When this is not possible, we use a multiple-comparisons procedure. This procedure determines which groups differ sig(cid:374)i ca(cid:374)tly (cid:449)he(cid:374)e(cid:448)e(cid:396) the o(cid:448)e(cid:396)all test rejects the null hypothesis that all groups have the same mean. Always accompany the anova by data analysis to see what kind of inequality is present. Plotting the data in all groups side by side is particularly helpful. Testing the equality of several means is helpful in understanding data.