Have an idea?

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

Point me to some helpful tutorials for data analysis

Hi all,

I'm new to conjoint analysis and sawtooth software but am attempting to learn it to support some medical patient preference work I'm doing.  I've created an ACA survey and piloted it with a handful of people - now I'm teaching myself the data analysis and I'm stuck. SSI web is great for collecting data but I'm not a big fan of SMRT so far.  I'd like to try exporting my raw data  into another program like spss or perhaps even excel and doing some analysis there so I can better teach myself the methodology.  I've gone through the SPSS CA tutorial but I'm not sure how to approach the exported ACA data in spss or excel.  I figure many people do this - can you point me to any tutorials/walkthroughs available I can look at.

asked Jan 3, 2012 by anonymous
retagged Mar 8, 2019 by Walter Williams

1 Answer

+1 vote
It's easy to get raw utilities out of ACA/SMRT and to SPSS.  Within SMRT, just go to Analysis + Run Manager, then highlight the run you'd like to export.  Then, click Export.  You are allowed to save the utilities in .csv format, which may be opened with SPSS.

While it is interesting to look at averages of utilities across respondents, and even to look at importances (although the importance computation has weaknesses), the best sorts of analysis with conjoint are typically done with Market Simulations, using a tool like SMRT (or a simulator that you've customized within Excel).

With market simulations, we set up a "base case", which typically contains multiple products in competition with another.  Then, we alter the test product, and see how the share of votes for it (share of preference) changes as we make those modifications.

For a good introductory article on Market Simulators, please see: https://www.sawtoothsoftware.com/download/techpap/introsim.pdf
answered Jan 3, 2012 by Bryan Orme Platinum Sawtooth Software, Inc. (198,515 points)
Thanks Bryan - that is a great resource.  It also made me aware of your book that it is derived from - its on my shopping list now.

I am playing around with the idea to use conjoint to look at preferences of patients for medical services and speculate they may segment into groups that share certain demographics such as medical history and age.  

My main holdup with using SMRT was that I could not figure out what happened to the demographic information collected along side the conjoint analysis.  I can export it fine out of SSS web but I do not see it when I convert it over to SMRT.  I assume its user error, and I need a quiet day to sit and read through the manual.  Do I need to import the demographics into SMRT a different way that the CA data?

Indeed, you get the segmentation data into SMRT as a second step.  First, you will have used Analysis + Run Manager to import the conjoint utilities into SMRT.

Then, prepare a .csv file that has the same Case IDs, plus any segmentation variables you want to move into SMRT for use in market simulations (as segments and weights).  

Within SMRT, use File + Merge Variables, and a little wizard appears that asks you to browse to that .csv file that contains the segments.  Once you've done that, you'll see those segmentation variables within SMRT when you are doing simulation runs and tabulating the utilities/importances within SMRT.