I have an independent variable grouped into irregularly sized classes (0, 1-5, 6-10, 11-20, >20), a continuous dependent variable, and several control variables.

So far I've been trying to stay close to the format of the observed data, treat the IV as categorical and run a One-way ANCOVA. Variances were unequal, so I opted for a mixed linear model instead. This however feels overly complicated for such a simple problem and I was considering to instead just transform the data into class midpoints (i.e., 0, 3, 8, 15.5, ?20?) and treat the IV as continuous to perform a simple regression instead. Would this be a valid approach? Are there any alternatives that might be more suitable?

Thanks in adcance


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.