In reading about quasibinomial regression:
The quasi-binomial distribution, while similar to the binomial distribution, has an extra parameter (limited to ||≤min{/,(1−)/} ) that attempts to describe additional variance in the data that cannot be explained by a Binomial distribution alone.
What is quasi-binomial distribution (in the context of GLM)?
In summary, if a model has over dispersion instead of binomial logistic regression we should use quasi-binomial logistic regression. Basically, if we are in doubt we can assume that over dispersion is present in the data.
https://medium.com/@mohsin.eee/quasi-binomial-logistic-regression-5cc50a8eb67f
A quasi-binomial model is used to model binary response data which exhibit extra-binomial variation.
https://www.scirp.org/journal/paperinformation.aspx?paperid=115008
Given that quasibinomial is more flexible than binomial regression in modeling and account for extra-binomial variation and overdispersion, what is the advantage to implementing binomial regression? Why not always go with the more flexible quasibinomial regression over the less flexible binomial regression? Is there a disadvantage to estimating quasibinomial model, as it seems that in general it’s described as being more flexible in modeling overdispersion but it is not described in terms of costs to doing so? Is there a risk of overfitting or diminished sensitivity?
