Have an idea?

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

Stat Testing in Excel Based Simulator

Hi,

I am looking to stat test share of preference between SKUs and within subgroups in an Excel based simulator, based on a volumetric CBC. I used the Volumetric CBC .CHS converter, resulting in volumetric weights. How do I account for the weights in the stat testing? I currently have a one-sample t-test as a placeholder with the below formula:

=ABS((Share1-Share2)/SQRT((Share1*Share2/n)+(Share1*Share2/n)))

Thanks in advance!

Lisa
asked Jan 16, 2018 by LisaVogl Bronze (530 points)

1 Answer

+1 vote
 
Best answer
Matched-sample t-tests would be a stronger test than the independent samples t-test you are showing.  But, they don't account for weighting.  As long as your weights across people didn't differ very much, I would probably just do matched sample t-tests and ignore the fact that I was using weights.

But, if you want to do a better job formally of accounting for the weights, you should follow the formulas like we do with the commercial Sawtooth Software simulator, described at:

http://www.sawtoothsoftware.com/help/lighthouse-studio/manual/index.html?effective_sample_size.html
answered Jan 17, 2018 by Bryan Orme Platinum Sawtooth Software, Inc. (159,785 points)
selected Jan 17, 2018 by LisaVogl
...