Physics Forums (http://www.physicsforums.com/index.php)
-   Set Theory, Logic, Probability, Statistics (http://www.physicsforums.com/forumdisplay.php?f=78)
-   -   Logistic Regression with Dummy Variables as regressors? (http://www.physicsforums.com/showthread.php?t=396750)

 maverick280857 Apr19-10 10:00 PM

Logistic Regression with Dummy Variables as regressors?

Hi

Is it possible to use both dummy variable regressors as well as categorical response variables in the same model? Consider the following model from cricket (you don't need to know the game to answer this question):

Y = mode of dismissal (0 = not out, 1 = bowled, 2 = caught, 3 = stumped, etc.) [categorical]
X1 = time spent at the crease [numerical]
X2 = balls faced [numerical]
X3 = position down the order (1 = opening, 2 = one-down, etc.) [categorical/dummy]

In this case can we use the logistic regression approach with the linear predictor modified to include one or more dummy variables?

The usual logistic regression setup consists of a categorical response and numerical regressors. I am asking if we can have 1 categorical response, 2 numerical regressors and 1 categorical regressor.

How does one analyze such a problem?