Subscribe Bookmark RSS Feed

Means displayed in a Nested analysis do not match Fit Y by X calculated means....why?

txnelson

Super User

Joined:

Jun 22, 2012

I am running a Fit Model where the effects are nested. The Means displayed in the Least Squares Means Table for the higher level effect (TT in my data), do not even come close to what a simple calculation of the means using Fit Y by X show.

12941_pastedImage_0.png

The description of the Least Squares Means Table indicates that the Mean displayed is the "sample mean for the given level".

12942_pastedImage_1.png

I would appreciate if someone more versed in Nested ANOVAs than I am can provide me with an explanation.

I have attached the data table with an embedded script.

Jim
1 ACCEPTED SOLUTION

Accepted Solutions
Solution

Jim,

Apparently it is the No Intercept option that you selected caused the problem. In other words, if you include the intercept, they match. I agree with you that means should match, but I don't know why this happens. Please contact support@jmp.com, and I am sure they will give you an answer :-)

3 REPLIES
ih

Community Trekker

Joined:

Sep 30, 2016

I am believe you are doing two different analyses here, the fit Y by X only looks at once X at a time, where the Fit Model is using TT and TN at the same time.  I ran the fit model using only TT and came up with the same means as the fit Y by X.

12946_pastedImage_0.png

12947_pastedImage_1.png

txnelson

Super User

Joined:

Jun 22, 2012

Isaac,

I understand that I am doing 2 different analyses, however, the documentation seems to indicate that what the Mean value should be is the Sample Mean, and that the Least Sq Mean will be different, not the Sample Mean.

Jim
Solution

Jim,

Apparently it is the No Intercept option that you selected caused the problem. In other words, if you include the intercept, they match. I agree with you that means should match, but I don't know why this happens. Please contact support@jmp.com, and I am sure they will give you an answer :-)