|
|
Specifying threshold values |
|
Message/Author |
|
|
Hi, I'm having some trouble grasping the use of threshold values when doing a monte carlo using ordinal variables. I'd like to generate data on a 7-point scale for a CFA of two factors with 5 items each, but every time I run the monte carlo simulation I get a data set that has only 0s and 6s. I've tried working from scratch as well as starting with the 5.10 mc example to no avail. I think I may be confused moving away from the binary example to my ordinal attempt; is there an example in the user guide that uses multi-point ordinal data rather than binary? Or, could you give me an example of what a line of thresholds might look like? Right now this is what I'm trying under model population: [u1a$1*.1 u1b$1*.1 u1c$1*.1 u1d$1*.1 u1e$1*.1 u2a$1*.1 u2b$1*.1 u2c$1*.1 u2d$1*.1 u2e$1*.1] (1); [u1a$2*.1 u1b$2*.1 u1c$2*.1 u1d$2*.1 u1e$2*.1 u2a$2*.1 u2b$2*.1 u2c$2*.1 u2d$2*.1 u2e$2*.1] (2); etc., through to ...u2e$6*1] (6); Thank you! Mike |
|
|
Please send the full output and your license number to support@statmodel.com. |
|
|
Hi, What was the resolution to this? I want to generate ordinal data for a CFA on a 5-point scale for one factor. Thank you! Melissa |
|
|
The user probably made an mistake that is not of general interest; we don't keep records of this from 7 years ago. You can check the examples in the UG on our website - for each example there is a corresponding Monte Carlo script posted. See, e.g., UG ex7.6. |
|
Back to top |
|
|