Two Questions About One-Way ANOVAs Dear Paul, Your questions are good ones. I'll try to answer each one right below it. Q.1 I'm trying to critique an article in another class in which several groups of people are compared on basis of their mean scores as a group on several tests. So, group A, for example, scored 7.5 on one test and Group b scored 8.5 on the same test and group C: 4.3 and so on. No correlations are done yet I am consistently given p-values on the comparisons and "F= 6.50" or something similar, which I guess is a calculated value of sorts. I also am occasionally given a "d" value. My core question is: Are comparisons like this as valid as correlations? A.1 After we tackle the content of Chapters 12 and 13, you'll know exactly what's going on in the article you're trying to critique. As for now, just consider the "F" test to be a way of comparing (simultaneously) the sample means from Groups A, B, and C to see if there's enough variability among them to cast doubt on the null hypothesis that all three corresponding population means are identical. Yes . . . they are valid comparisons! Q.2 Any clues on how to tell the practical from statistically significant? How can I interpret this with present knowledge or do I need to get further in the book? Do I need to bring the article by? A.2 The issue of practical significance, in my opinion, is mainly a matter for people in a discipline (and especially for those who work "in the field" or "in the trenches") to determine. There are a few statistical procedures that give the appearance of helping . . . but I honestly think that a practitioner, using no statistical tools at all, is the expert whose opinion should be listened to AND TRUSTED. Sky Huck