"this parameter is set to zero because it is redundant" - SPSS Ordinal Regression output My question has been asked here but I was none the wiser with the short answer given - 
Generalized Linear Model on SPSS with the 'error': "set to zero because this parameter is redundant"
This "parameter is set to zero because it is redundant" is appearing on the output and is always the last of all my my factors (e.g. job position 1-6, it says it for the 6th one, region of work 1-3, it says it for the 3rd). I have tried switching the numbers around, so work 3 becomes work 1, but then work 2 (now 3) says "parameter is set to zero because it is redundant".
Any help as to what I can do would be really appreciated. I have looked on a lot of different internet pages to no avail. Simple explanations really appreciated.
Many thanks.
 A: Suspicion
From your story, I assume you have dummy-coded job positions like this:
PersonID   Job1   Job2   Job3   ... 
1          0      1      0      ...
2          1      0      0      ...
..         ..     ..     ..     ...
100        1      0      0      ...

Description
In a normal regression model, you will have the intercept of your regression equation. For mathematical reasons, your regression can only come up with coefficients, when all variables in the data present add 'unique' information. Else you will suffer from perfect multicollinearity, in your case there is a special name for it: the dummy trap. 
I'll discuss briefly: The regression includes an intercept, usually called $\beta_0$. If you also include a dummy variable for every job position, the predicted outcome for every person will be
$\hat{y} = \beta_0 + \beta_1D_{job1} + \beta_2D_{job2} + \beta_3D_{job3}$ 
So you have three groups of people in your data (job1, job2 and job3), one job per person. 


*

*For a given person in job1, the $\hat{y} = \beta_{0} + \beta_1\cdot
   1$ 

*For a given person in job2, the $\hat{y} = \beta_{0} + \beta_2\cdot
   1$ 

*For a given person in job3, the $\hat{y} = \beta_{0} + \beta_3\cdot
   1$


The problem is that the regression will not be able to separate the magnitude of $\beta_0$ from the other coefficients. SPSS lets you know this error occurred by means of the error message you see. It automatically cancels one of your variables (see option 2 below).
Solution
There are two options that will solve this problem (option 1 is easiest):


*

*Run a regression without an intercept 


To fit the model without an intercept, one needs to use Analyze > Regression > Linear, click Options (after specifying the model), and then uncheck the Include constant in equation box.


*Remove the variable that indicates job3 (or job2 or job1). $\beta_{0}$ will now be the coefficient for the job you removed (say we remove job1). The other coefficients change their interpretation in the sense that they represent the difference between job1's coefficient and their own coefficient.

