|
|
Non-convergence with centered data |
|
Message/Author |
|
Tabea Smith posted on Sunday, December 14, 2008 - 12:31 pm
|
|
|
I'm running a 3 variable latent difference score model. When I center each variable manually (i.e. removing the mean), the models do not converge. However, when I use: CENTERING is grandmean (y1 y2 y3 y4 y5 x1 x2 x3 x4 x5 z1 z2 z3 z4 z5); the models converge. Shouldn't these methods be identical? I've checked to make sure that the values for each variable are not too small (to rule out precision issues). As an aside question, I'm wondering why the estimated sample statistics from TYPE=BASIC show means like -.03, -.04, etc. Thank you. |
|
|
Will need to see your three runs. Please send input, output, data and license number to support@statmodel.com. |
|
Back to top |
|
|