Have an idea?

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

Merging similar MaxDiff studies

I created a survey with 3 maxdiff studies: each have the same items (8), items per subset (4)and same number of subsets shown to participants (5). I needed 3 different ways to pose a question to different groups of people. I generated 3 different designs for these studies and I have a feeling this was wrong after reading the forum posts but I have collected data now.

Can I merge this data to use with the latent class software?
If so, would I just convert my CHO files I get from SSI into CSV files and then merge the data into one CSV file? The respondent identification is fine as the 3 maxdiffs are within the same survey.

Thanks a lot,
Paul Balcombe
related to an answer for: How to merge data for a Maxdiff
asked Mar 26, 2013 by anonymous

1 Answer

0 votes
If the items were the same, as well as the items per set, and numbers of sets, then that's good.  But, I don't know what you mean when you say "I generated 3 different designs for these studies".  It probably means you had three separate studies, and used the default settings which would generate 300 "versions" per study.  Each study probably used the same starting seed?  So, in reality these studies all used the same 300 designs?

Now, if you just used a single version (block) of the design for each unique group of respondents, this would be a bad thing, as you would confound pychological order/position effects with true differences in utility between groups.  But, I doubt this is what you did.

So, assuming you used lots of different versions (e.g. dozens or hundreds), even if they were generated using different design seeds, just combine them all within one .CHO and submit to latent class software.  No harm done.
answered Mar 26, 2013 by Bryan Orme Platinum Sawtooth Software, Inc. (160,485 points)
Yes I created 3 separate studies, with the same default settings (300 versions and design seed: 1), so I have used lots of versions.

So I can combine them within one CHO file. I'm not sure how I can actually combine the 3 CHO files or create one merged file from the software. I have unsuccessfully played around with combining the files but I can't find a description on how I could do this.

Thanks,
Paul
...