Have an idea?

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

Insignificant from T-ratio but why the attribute importance is high?

I did Latent Class analysis, in one group..the T-ratio from all the level of an attribute turns to be insignificant. However, the attribute importance of that level is at a high number  (2nd rank when comparing with other groups). I wonder why?

Also, does this means once it turns to be significant.. the preferred attribute level and attribute importance cannot be interpret?

Thank you so much!
asked Jan 5, 2020 by Emily

1 Answer

+1 vote

Hmmm, can you tell me more?  The most obvious culprit if you're running this as an aggregate logit model (LC with one class) might be a small sample size - how many respondents do you have?

In any case, it certainly does not mean what your second paragraph suggests.
answered Jan 5, 2020 by Keith Chrzan Platinum Sawtooth Software, Inc. (99,300 points)
I did LCA and choose 5 segments.Sameple size is 185

So, this means I can still interpret the  preferred attribute level and attribute importance but it just not strong argument as it's not significant right?
Thank you Keith!

Yes, splitting a sample of 185 people into segments that average 37 people, non-significance will be a common occurrence.