For a simple Poisson regression, I would interpret the coefficient as

(1-exp(coef))*100 for percentage change in y given a unit change in x

So for a Conway-Maxwell Poisson regression, Sellers and Shmueli (2010) suggested a crude comparison of dividing the coefficients by the shape parameter 'v' or 'nu'. In this case would the marginal effect just be (1-exp(coef/nu))*100?

  • $\begingroup$ Sorry for the confusion I think I have missed my point of the question, I wanted to ask about the percentage change in y given a unit change in x. This will imply exp(coef) to obtain the multiplicative factor than subtract this number from 1 and multiply by 100 to get the percentage change. However I am not sure if this apply to the Conway Maxwell regression. I shall edit the question. $\endgroup$ – Michael Qiu Mar 10 at 14:08

Your Answer

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

Browse other questions tagged or ask your own question.