Linear Regression in Mplus vs SAS PreviousNext
Mplus Discussion > Structural Equation Modeling >
Message/Author
 Annonymous posted on Tuesday, January 17, 2006 - 9:56 am
Hi - am new to Mplus, and as a way of 'double checking' that my coding is correct, I am running the same models in SAS.
SAS: A = B for the regression of B on A

MPLUS: B on A (in the hopes that I get the same results from SAS)

A is continuous, B is ordinal.

the values are quite different - is this due to differences in the estimation method or to incorrect coding in Mplus on my part?

also - if i wanted to adjust the relationship between A and B for the influence of C, how would that coding look?

thanks!
 bmuthen posted on Tuesday, January 17, 2006 - 10:51 am
I believe saying A = B in SAS means that B is the predictor of A (B is x and A is y). So this would translate to A ON B in Mplus.

I don't know what SAS does with an ordinal predictor. Mplus treats it as continuous unless you want to view it as an nominal variable in which case you split it up into 0/1 dummy variables.
 Annonymous posted on Tuesday, January 17, 2006 - 10:57 am
Ok - this seems to help. I"m getting the same values from MPlus when I say A on B.

However, in trying to adjust for a third variable, the values for the third variable
are still very different, even when C is a continuous variable. The coding that I've used is
A on B C

is this correct? any thoughts on the difference, given that both programs get the same value for B?
 bmuthen posted on Tuesday, January 17, 2006 - 11:08 am
That should give the same results, so there is probably something else going on such as not reading the data the same way. To diagnose most effectively, please send the input, output, data, and license number to support@statmodel.com.
Back to top
Add Your Message Here
Post:
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Password:
Options: Enable HTML code in message
Automatically activate URLs in message
Action: