Chi Squared vs. F Test

D

Dolomite

New member
Joined
Jun 18, 2026
Messages
0
Reaction score
0
What's the difference, aren't they both measuring the difference in variances of independent variables b/t the normal distribution of a population?

thanks..
 
One more question: Does anyone know if the schweser practice exams book is the same level of difficulty as the sample exam workbook?

thanks
 
Dolomite Wrote:
-------------------------------------------------------
> What's the difference, aren't they both measuring
> the difference in variances of independent
> variables b/t the normal distribution of a
> population?
>
> thanks..

Chi square test stat is a test of hypthesis about the population variance

F-test is comparing two variances (similar to paired analysis)
 
"One more question: Does anyone know if the schweser practice exams book is the same level of difficulty as the sample exam workbook? "

I think schweser practice is harder.
 
oatmeals2 Wrote:
>
> Chi square test stat is a test of hypthesis about
> the population variance
>
> F-test is comparing two variances (similar to
> paired analysis)


And these two tests have one thing in common: they are all applied to interval data.
 
I don't buy that last part.

Chi-square and F are distributions. There are lots of hypothesis tests that use them. A chi-square is the sum of a bunch of squared normals and an F is a multiple of the ratio of two chi-square distributions.

Probably the most used chi-square test is the contingency table chi square which is nominal data. An ANOVA F-test is based at least on nominal independent variables.
 
Hi, Joey,
I meant: when chi-Squared test is used to describe a population variability, as oatmeals2 mentioned, it has to be interval data. You must talk about when we compare two or more populations, the chi-squared goodness-of-fit test and chi-squared test of a contingency table, in that case, we do deal with nominal data.

F-test is used to compare two variances, and ANOVA table, in which the results of the analysis of variance reported, is dealing with interval data too, since nominal data don�t even have a center. Where did you get your idea from? just curious...
 
Suppose I want to compare the CFA test scores of Democrats, Republiicans, and Independents. I get the data from CFA institute and do plain vanilla one-way ANOVA. My independent variable is political party, which is nominal.

Here's an interesting idea that hasn't helped me much in life but someday it's going to. The measurement characteristics of data fall on a continuum (not just ratio, interval, etc), their measurement characteristics are empirical, the measurement characteristics depend on the test used (not the reverse as is usually thought). One of my grad school professors told me that, didn't explain it, and I have been pondering it for 20 years. I don't think it's utter BS.
 
You must log in or register to reply here.
Back
Top