"G.S.Clarke" wrote: > First � is there any good reason why kruskal-wallis (which I would have > taken as being the more powerful technique) doesn�t show significance > when the chi-squared tests do (in other words, am I correct in my > assumptions as to which is the more powerful technique).
The data and hypotheses are different so power is hard to compare. In this case my guess is that KW is not very powerful when there are lots of ties. > Secondly chi-squared provides a �symmetrical� analysis � it says (at > least I think that it does) that having information about one of the > variables provides information about the other variable � and vice > versa. However is Kruskal-Wallis �symmetrical� in this sense. Does not > KW say (had it been significant) that the marital status of the subjects > (the factor) informs us somewhat on their educational status (the > dependent variable). Now logically if this is true then the educational > status should inform us somewhat on the marital status, however is this > also the case �statistically� i.e. will any significance level generated > by the KW using education (dependent) v marital status (independent) > also apply to phrasing the question the �other way around�? I think it is symmetrical - but note that the educational rank of the groups are compared - not the educational category (the data are not really the same). For example, Chi-square will probably be more sensitive to concentration of one educational category in one marital category. If that concentration produces a decrease in another category the mean ranks of the groups might not differ. e.g., simple example ... E1 = no education, E5 = postgrad, A = married, B = unmarried category A: E1, E1, E1, E5, E5, E5 category B: E3, E3, E3, E3, E3, E3 Mean ranks don't differ, but the category associations do! Thom . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
