Share your ideas for the JMP Scripting Unsession at Discovery Summit by September 17th. We hope to see you there!
Choose Language Hide Translation Bar
Highlighted
jmpvmd
Level II

How to determine differences between each pair of categorical data with 3+ groups

Hi!

 

I feel my question is pretty straight forward. How can I determine differences between each pair of categorical data with 3+ groups?

I have two categorical variables:

1) Type of Grant (Grant 1, Grant 2, Grant 3, & Grant 4)

2) Professorship (Assistant professor, associate professor, full professor)

 

I can use Fit Y by X and get a significant Chi Square value between the two.... But how can I tell where this differences lies? For example, how do I know if grant 3 is significantly more likely to have associate professors compared to grant 2, or that grant 1 is significantly more likely to have full professors than all of the other grants?

 

If I make professorship a numerical variable, I can get comparisons and a p value for each pair, is this how it should be done?

 

Thanks in advance.

1 ACCEPTED SOLUTION

Accepted Solutions
Highlighted

Re: How to determine differences between each pair of categorical data with 3+ groups

Click the red triangle next to Contingency and select Correspondence Analysis.You only have two variables so it reduces to a one-dimensional plot. You will see the association between levels.

 

Here is mock data to illustrate.

 

Screen Shot 2020-06-28 at 8.52.16 AM.png

 

Here is the initial analysis as you describe.

 

Screen Shot 2020-06-28 at 8.52.50 AM.png

 

Here is the addition of the correspondence analysis.

 

Screen Shot 2020-06-28 at 8.53.04 AM.png

 

So in this mock up, Assistant is associated with Grants 1 and 2, Associate is associated with Grant 4, and Full is associated with Grant 3.

 

You can read more about correspondence analysis in Help > JMP Documentation Library > Basic Analysis guide > Contingency Analysis chapter.

 

Note that there is a richer platform, Analyze > Multivariate Methods > Multiple Correspondence Analysis available. It is described in Help > JMP Documentation Library > Multivariate Methods guide > Multiple Correspondence Analysis chapter

Learn it once, use it forever!

View solution in original post

1 REPLY 1
Highlighted

Re: How to determine differences between each pair of categorical data with 3+ groups

Click the red triangle next to Contingency and select Correspondence Analysis.You only have two variables so it reduces to a one-dimensional plot. You will see the association between levels.

 

Here is mock data to illustrate.

 

Screen Shot 2020-06-28 at 8.52.16 AM.png

 

Here is the initial analysis as you describe.

 

Screen Shot 2020-06-28 at 8.52.50 AM.png

 

Here is the addition of the correspondence analysis.

 

Screen Shot 2020-06-28 at 8.53.04 AM.png

 

So in this mock up, Assistant is associated with Grants 1 and 2, Associate is associated with Grant 4, and Full is associated with Grant 3.

 

You can read more about correspondence analysis in Help > JMP Documentation Library > Basic Analysis guide > Contingency Analysis chapter.

 

Note that there is a richer platform, Analyze > Multivariate Methods > Multiple Correspondence Analysis available. It is described in Help > JMP Documentation Library > Multivariate Methods guide > Multiple Correspondence Analysis chapter

Learn it once, use it forever!

View solution in original post

Article Labels