Sure see the regression form of ANCOVA,. covered here:

http://www.stat.cmu.edu/~hseltman/309/Book/chapter10.pdf

D. Booth

Dear Dr. Vederhus

You can check my article:

Article A Bayesian Logistic Regression approach in Asthma Persistenc...

A dummy variable is encoded as follows:

If you have three categories let's say A,B,C then you create three different new variables (dummy).

The first contains 1s in the position you have As and 0s to all the others. The second contains 1s in the position you have Bs and 0s to all the others. And the third contains 1s in the position you have Cs and 0s to all the others.

Of course one of them can be ignored as its information is contained in the other two.

To understand better this example you can check my article at the METHODS section (Clinical data subsection).

John-Kåre Vederhus , what software are you using?

SPSS, I suppose the analysis is rather straightforward in all the software available? I just wanted article suggestions where it has been described.

jk

See also the attached. Best, D. Booth

Hi John. With SPSS, you can use UNIANOVA rather than REGRESSION to estimate your model. (If you do estimate the model both ways, you'll find that you get the same model R2.)

• https://www.ibm.com/support/knowledgecenter/en/SSLVMB_26.0.0/statistics_reference_project_ddita/spss/base/syn_unianova.html

With UNIANOVA, there is no need to compute dummy variables. Simply include categorical variables as "fixed factors". You may also find the CONTRAST and EMMEANS options helpful. HTH.

Thanks, Bruce Weaver . I suppose you mean the "Univariate" choice from "General Linear Model" menu?

Best

John-Kåre

That's right, John. Analyze > General Linear Model > Univariate.