Optimized Hot Spot Analysis

1443
3
02-16-2014 07:31 AM
PinkOrchid
New Contributor
Hello everyone,

I have a dataset of respondents (a point dataset) and their responses on a customer satisfaction scale with low, medium, and high categories. I would like to determine hotspots based on their satisfaction. Since Getis and Ord's Gi* test is based on the assumption of normality and my variable is not normally distributed, what test can I use to determine the potential statistically significant hotspots.

Thank you very much!
Best regards.
0 Kudos
3 Replies
SedaKahraman1
New Contributor
Hello everyone,

I have a dataset of respondents (a point dataset) and their responses on a customer satisfaction scale with low, medium, and high categories. I would like to determine hotspots based on their satisfaction. Since Getis and Ord's Gi* test is based on the assumption of normality and my variable is not normally distributed, what test can I use to determine the potential statistically significant hotspots.

Thank you very much!
Best regards.



Hi,

As I know, regression analysis require normally distributed data. Getis and Ord's Gi*  metod interests  in   only p-values, z score, and gi-index. So you can determine that your values significant or not using  p-values, z score, and gi-index.
0 Kudos
PinkOrchid
New Contributor
Hi,

As I know, regression analysis require normally distributed data. Getis and Ord's Gi*  metod interests  in   only p-values, z score, and gi-index. So you can determine that your values significant or not using  p-values, z score, and gi-index.


Hi Seda,

Thank you so much for your answer! I am still not sure if I can do that just to test whether the data is random or it shows some pattern- I would think so, but I need some proof:). The textbooks I found seem very relaxed in describing these things. I am surprised I was not able to find too much detail... I am still searching for an answer to this...

Thanks again!
0 Kudos
SedaKahraman1
New Contributor
Hi Pink,

Would you send me these texboox please(maybe link),  I have never heard this issue.
0 Kudos