I have a question about my MaxDiff analysis.
I have created a questionnaire with about 1,500 participants. I have performed the latent class analysis and the HB analysis.
The results of the latent class analysis show that I can create up to 4 groups that I can interpret. For the first three groups, the results are fairly clear and easy to interpret, while for the fourth group, the rescaled values are fairly evenly distributed.
So I checked the fourth group with the fit statistic (correspond to the RLH- value?) I got from the HB analysis (does that make sense?) and I saw that the participants in this group tend to have a very low score (around 0.25 - 0.35), while they have scores between 0.35- 0.80 in the other groups.
So my question is whether I should eliminate the people with low fit stat and try another run, and if so, how can I do that, because when I eliminate them in the exported Excel file and then try to re-upload the file for analysis, I always get the following error message:
"an invalid field 'sys_pagetime_29;;;;'. ERROR: The name contains an invalid character."
Thank you in advance for the help!!