topic Re: Doing a Oneway Analysis with Multiple Response Columns in Discussions

Doing a Oneway Analysis with Multiple Response Columns

Earendur — Sat, 10 Jun 2023 23:26:20 GMT

Below is a sample data table that shows what I am trying to achieve. I want to compare means of serum levels based on a patient's mutations. The problem is some patients have multiple mutations in the same gene.

I would like JMP to recognize that the Mut1, Mut2 need to be assigned with the other mutations with the same name instead of creating its own category for them. I have tried setting the column to "Multiple Response" but then the Fit Y by X platform will not accept "Gene 1" as an X factor.

Others have suggested creating a second column for the extra mutations but then the Fit Y by X platform creates two charts and I can't analyze the means between the different mutations.

I've looked and looked for a solution, I know there has to be something simple I am missing.

Re: Doing a Oneway Analysis with Multiple Response Columns

statman — Sat, 20 Feb 2021 19:52:24 GMT

I'm not sure I have an answer, but the way your table is laid out, you have a comma between the two mutations Mut1, Mut2. Why not identify it as Mut12 or MutAB, etc.

Re: Doing a Oneway Analysis with Multiple Response Columns

txnelson — Sat, 20 Feb 2021 19:58:05 GMT

One way of handling this is to replicate the Serum data for each mutation and then run the Oneway analysis. This is easy to do in JMP.

Click on the header for Gene 1
Select from the pull down menu Cols=>Utilities=>Text to Columns
Enter a ',' as the delimiter and select OK
Now the data needs to be stacked so select from the pull down menu Tables=>Stack
Select both Gene 1 1 and Gene 1 2 as the Stack Columns
Click on OK
Now you can run your Fit Y by X, specifying
1. The column called Data ax your X Factor
2. The column Serum Level Y and your Y Response
3. Click on OK