Have an idea?

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

Significance between attribute levels


I am running a ACA survey and downloaded my data now to analyse with the SMRT. I already estimated my raw data with OLS. So I have the average utilities of each attribute level and average importance plus the standard deviation in Excel to have an overview.
Now I would like(I assume with a t-test)  to check the differences between attributes and the differences between the attribute-levels is significant.
Could you explain to me if there is a function in SMRT to do so?

I would also like compare groups so that I can say that for example femal and male has this or that utilities/preference for a designed product. How can I measure the significance between these groups? What do I need to click in the market simulator? Do I need to calculate "share of preference" ?
I already found this formular from the conference in 2004 - when I would like to check the significance difference for products and simulating a purchase likelihood - can I use the same formula - just replace "share" to "purchase likelihood" from female or rather male? Does that work to test the significance?

Thank you very much for your help!

asked Feb 10, 2013 by anonymous
retagged Feb 10, 2013 by Walter Williams

1 Answer

0 votes
Our newsletter article covering these issues is at: https://www.sawtoothsoftware.com/education/ss/ss20.shtml#ss20sig

There isn't an automatic function in SMRT to do this...but you'll see that SMRT allows you to export the individual-level utilities or individual-level shares of preference that allow you to make such computations in Excel.
answered Feb 11, 2013 by Bryan Orme Platinum Sawtooth Software, Inc. (134,015 points)