ANOVA (Several means test) Bonferroni Correction, and finding F critical values with significance levels other than .05 and .01

Question

I understand how to perform a one factor ANOVA test for several means. However, Ive tried learning about Post-Hoc Comparisons and it seems to me that they compare every possible pair of groups individually to determine which groups are the ones that differ. We use the Bonferroni Correction alpha/tests in our university. This raises a couple of questions for me. With this new alpha, we will most likely not have significance levels of .05 or .01 or other numbers on a t-distribution chart. How then, are we supposed to find the critical values? Is the standard procedure just listing inequalities of the two closest critical values? i.e., df = 4 and alpha calculated with bonferroni is .17, thus 3.747 < tc < 4.604?

Answer 1

I assume you meant that the alpha associated eith the Bonferroni in your example was 0.017. However, that is just a nitpick. In answer to your question, yes you could use the inequatities. Or, alternatively, estimate the range of the p-value for your calculated t-value for the comparison. If you are allowed to use a graphiing calculator or other statistical packages, they willl allow you to calculate the specific t critical value. Also, many statistical packages, such as SPSS, will provide you with several options for the multiple comparisons.