In simple logistic regression, a common interpretation of the model coefficient $\beta$ is that a 'one-unit increase in the independent variable is associated with an increase of the log-odds ratio of the outcome variable by $\beta$'). This is already answered here.

What confused me is that the logit function is not linear with respect to $x$, so how can a one-unit increase in $x$ lead to a constant change in the log-odds ratio?


1 Answer 1


In logistic regression we have that:

$$ \ln\left({\frac{p}{1-p}}\right) = \mathbf{Xb} $$

where $\mathbf{Xb}$ is the linear predictor, $\mathbf{X}$ being the model matrix of explanatory variables and $\mathbf{b}$ the vector of coefficients. $\ln\left({\frac{p}{1-p}}\right)$ is the logit function (log of the odds).

Therefore every component of $\mathbf{X}$ (eg, an individual explanatory variable) is proportional to the log odds and so any change in one of them has a linear effect on it.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.