Big Class: Creating a New Grouping Variable

Question

Hi! I'm sure this will be elementary for many of you, but having come from tabulation software to JMP, this one is stumping me.

I am trying to understand the process of creating a new grouping variable. This is a relatively straight forward process in IBM's Survey Reporter, but doing so in JMP is not immediately clear to me.

Using the Big Class data file, I would like to create a new column where the grouping variable will reside. I would like this variable to capture all Males that are 13 years old.

Could someone please walk me through the process of creating this, or point me in the direction of some documentation that reviews this process?

melinda_thielba · Answer

In your case, you probably want to use Recode. Right-click the column name in the middle left pane of the data table and select Recode.

You can use the resulting dialog change everything below 13 to a 1 (or whatever value you prefer), 14 - 15 to a 2, etc.

The In Place button at the upper right means the default is to recode the data "In place" (i.e. replacing the current values). If you select that button, you will also have the option of creating a new column (which means you would have two age columns after you click OK, one that is grouped and one that is not), or creating a Formula column.

The formula column will also be a second Age column, but formula columns update themselves if there are changes to the original column. So, if you created a column called Grouped Age that was a formula column based on Age, and then realized that one of the original ages was miscoded, all you would have to do is fix the original value. Grouped Age would update.

These are the instructions for JMP 11. Recode has been enhanced in JMP 12. You can read about the new features in this blog post: Coming in JMP 12: Overhauled Recode command

Please reply if you need more information or have further questions!

pmroz · Answer

If you want to search for Males that are 13 years old, click on Rows > Row selection > Select where

Select the age column and then put the value of 13 in the text box.  Click on Add Condition

Select the sex column and put the value M in the text box.  Click on Add Condition.

Click OK

That will highlight the rows having males that are 13 years old.

On the other hand if you truly want a grouping variable you could create a new column having the formula

:sex || char(:age)

Then the rows you want are indicated by M13