cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 
Browse apps to extend the software in the new JMP Marketplace
Choose Language Hide Translation Bar
Rongyu_Kuang
Level I

How to compare other two coefficients except the coefficient from the reference group in a regression model with dummy variable

In a regression model with dummy variables: 

 

For example, I would like to test the relationships between occupation types and salary. In this condition, occupation is a categorical variable with 4 types (i.e., occ1; occ2; occ3 and occ4); salary is a continuous variable. 

 

In the regression model to test the relationships between occupation types and salary, I regard occ4 as the reference group, valued occ4 as "0", and valued the other three occupation types (occ1, occ2, and occ3)  as "1" respectively. To sum up, in this regression model, I got three dummy variables (i.e., occ1=1/occ2=1/occ3=1). 

 

In this vein, I run the regression model with three dummy variables and got the regression formula like: Y(salary)= B0+Bocc1X+Bocc2X+Bocc3X+error. In this model, Bocc1, Bocc2 and Bocc3 is calculated by comparing to the referent group (i.e., occ4). However, if I would like to compare two coefficients (e.g., Bocc1 and Bocc2), how should I implement it through JMP? Should I use t-test? but how to achieve it? 

 

Many thanks indeed for your help. 

1 REPLY 1

Re: How to compare other two coefficients except the coefficient from the reference group in a regression model with dummy variable

JMP automatically handles the coding for you so you do not have to create dummy variables. Add the Value Order column property to the Occupation data column and make the reference occupation (occ4) the last level. Set up the analysis by selecting Analyze > Fit Model. This dialog will launch the Fit Least Squares platform by default. Then see how to use Estimates > Custom Test or Multiple Comparisons from the platform red triangle menu.