Discrepancies in Estimated Cox PH Coe... PreviousNext
Mplus Discussion > Growth Modeling of Longitudinal Data >
 Michael Arnold posted on Thursday, November 22, 2007 - 1:10 pm
I am exploring the role of neighborhood characteristics and individual observed and latent factors on 5-year survival among persons living with AIDS. Mplus seems to be ideal for this.

I ran a very simple Cox proportional hazard model on (0,1) dummy variables of race/ethnicity i.e. MODEL: aids ON blk lat asn), and the Mplus estimates are very different from those I obtain using the cox approach in R (and are not in line with the K-P plots).

*The R PH output is as follows:
race coef exp(coef) se z p
blk 0.292 1.339 0.107 2.72 0.0066
lat -0.241 0.786 0.142 -1.70 0.0890
asn -0.310 0.734 0.249 -1.24 0.2100

**By contrast, the Mplus estimates are:
race Est SE Est/SE
BLK 0.008 0.010 0.808
LAT 0.005 0.010 0.520
ASN 0.052 0.025 2.098

Can you please help me understand why I am returning such different results?
 Linda K. Muthen posted on Friday, November 23, 2007 - 10:03 am
If you are following Example 6.21, you should get the same results. You may have something specified incorrectly. Please send your input, data, output, and license number to support@statmodel.com. Please also send the R input and ouptut.
Back to top
Add Your Message Here
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Options: Enable HTML code in message
Automatically activate URLs in message