I have data with a large proportion of zeros that we are committed to analysing with ITS [to fit in with other parts of a project] and have been advised that the best approach would be to convert to a binary outcome. I've done so and fit a binomial glm to the data, but[with experience of more standard count/rate based ITS analyses] I am not too sure of whether yhe outcome should be reported as the coefficients of the model as normal? I u derstand that such would be an Odds Ratio, but wondered whether the coefficients were also less relevant?
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
If I understand your question correctly you fit a binomial (binary outcome) regression, if this is true then I think the coefficients to be interpreted as Odds Ratio. https://stats.oarc.ucla.edu/other/mult-pkg/faq/general/faq-how-do-i-interpret-odds-ratios-in-logistic-regression/ Hope this answers your question.
