Fit Y by X Nominal Variables

[#1]

Hi All,

I'm having trouble using Fit Y by X.

It works when I used a this one particular Y variable, which is 2 levels (e.g., 'yes', 'no'), character type, nominal. The X variables I've used are also character type, nominal, but range from 2-5 levels (e.g., 'increase', 'decrease', 'no change').

But when I use a different Y variable, also 2 levels (e.g., 'low', 'high'), character type, nominal, it does not work. I enter the variables into Fit Y by X and get a blank screen. 

I checked the Column Properties and there isn't anything obviously different between the working Y variable and the non-working ones.

Any suggestions or help would be great!

Minna

[#2]

If you are getting a blank screen similar to the one below, it is because none of the rows in your data have values for both of the columns in your analysis.

[#3]

This is the blank screen I'm getting:

And here is an example of it working with a different Y character/nominal variable:

[#4]

It seems as though there is an issue with the data. Is there any way that you can share the data or an anonymized version of it?

[#5]

Hi Dan,

As part of de-identifying the data, I shortened the variable/column names. I test-ran it to reproduce the same blank screen, but didn't. Instead, it worked! 

The one variable that worked previously was a short name. It would only work, though, if it was entered as a Y variable in Fit Y by X. But it would produce just a blank screen if entered as an X variable.

But now all my categorical/nominal variables - with shortened names - work as either the Y or X variable. Strange, no?

Either way, thank you!

Minna

[#6 · Accepted solution]

First, I am glad that you are getting your analysis. However, this behavior does not sound right. I would strongly encourage you to contact the JMP tech support team regarding this issue. They can be emailed at support@jmp.com

[#7]

Will do, Dan. Thanks again!