topic Re: Dependent Variable of Logistic Regression in Statistical Procedures

Dependent Variable of Logistic Regression

deleted_user — Tue, 23 Nov 2010 20:21:32 GMT

Can the dependent variable be continuous but between 0 and 1? Any publications on this issue?

Re: Dependent Variable of Logistic Regression

Paige — Tue, 23 Nov 2010 20:31:56 GMT

I don't think this is logistic any more. It sounds like a generalized linear model, and you would need to know the distribution of the dependent variable errors.

Re: Dependent Variable of Logistic Regression

Ksharp — Wed, 24 Nov 2010 04:25:17 GMT

That is impossible.If dependent variable is continuous,that is not necessary to use logistic distribution to map discrete value to infinite(negative and positive ).If dependent variable is continuous ,GLM is suited for that.

The appearance of logistic model is just to solve the problem of discrete value of dependent variable.

Ksharp

Re: Dependent Variable of Logistic Regression

StatDave — Thu, 02 Dec 2010 20:27:54 GMT

While logistic regression involves a binomial or multinomial response, it is still reasonable to fit a logistic curve to a continuous response that ranges from 0 to 1. This can be done by fitting a nonlinear model such as p=1/(1+exp(-xbeta)), where xbeta is a linear function of your predictors. Also, Collett (2003, pp 329-330) suggests a quasi-likelihood approach that fortunately is similar to using events/trials syntax in PROC LOGISTIC with the proportions as the events variable, with the trials variable equal to 1, and with the SCALE=DEVIANCE option in the MODEL statement.