Contents
What is fractional response model?
Fractional response models are for use when the denominator is unknown. Fractional response estimators fit models on continuous zero to one data using probit, logit, heteroskedastic probit, and beta regression. Beta regression can be used only when the endpoints zero and one are excluded.
When performing logistic regression we are trying to?
Logistic regression analysis is used to examine the association of (categorical or continuous) independent variable(s) with one dichotomous dependent variable. This is in contrast to linear regression analysis in which the dependent variable is a continuous variable.
Which is the linear predictor in fractional regression?
In the following, y is our target variable, X β is the linear predictor, and g (.) is the link function, for example, the logit. As such, we can just use glm like we would for count or binary outcomes. It will warn you that the outcome isn’t integer as it expects, but in this case we can just ignore the warning.
Are there robust standard errors in fractional regression?
Also, as noted in the StackExchange link in the references, while by default the variance estimate is ‘robust’, possibly leading to standard errors that are similar, the basic result is not the same as using the robust standard errors. We can get robust standard errors for the quasi-likelihood approach as well, but they were already pretty close.
Which is the binomial likelihood for fractional regression?
It turns out that the underlying likelihood for fractional regression in Stata is the same as the standard binomial likelihood we would use for binary or count/proportional outcomes. In the following, y is our target variable, X β is the linear predictor, and g (.) is the link function, for example, the logit.
Are there any Stata commands that allow fractional response?
fracglm fills gaps left by other Stata commands. Commands like logit, probit and hetprob do not allow for fractional response variables. glm can estimate some fractional response models but does not allow an equation for heteroskedasticity.