Have an idea?

Visit Sawtooth Software Feedback to share your ideas on how we can improve our products.

respondent weight in latent class

hi

I'm setting up a maxdiff segmentation in latent class, where i have for each respondent a sample weight (sociodemo based weight).
I've modified the cho file to inlcude the respondent weight as such

3857 1 16 26 0
1,40601213620257
4 1
 0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  1
 0  0  0  0  0  0  0  0  1  0  0  0  0  0  0  0


and added a .wt file for analysis in LCA V2. the par file etc all look ok, but i get an error when running stating my CHO file in incomplete. thought it might be regional settings, so also tried with . as decimal seperator, but still the same.

tx for the help
asked Apr 16, 2014 by frederik

1 Answer

0 votes
That version of Latent Class is very old and does not support demographic variables (which you're using as the weight) with decimal numbers.  The current version (v4.7.0) supports this feature.

The .cho file also requires a period as a decimal (the format is not locale-aware).
answered Apr 16, 2014 by Walter Williams Gold Sawtooth Software, Inc. (15,075 points)
...