Logistic regression with an independent variable that is a proportion

[Sarah Haynes]

Hello,

I apologize if this is a simple question but I cannot find the answer in previous posts. In my logistic regression model, one of the covariates is a proportion (the proportion of a liquid made up of a certain chemical). If I wanted to interpret the odds ratio for this variable, is there something I need to do to transform it?

Any advice very welcome.

Thank you,

Sarah

[Maarten Buis]

Probably multiply that variable by 10 or 100 or 1000. If you don't a unit change in that variable represents going from nothing to everything, which is usually not a realistic or meaningful change. If you multiply by 10 the unit is a 10 percentage point change, if you multiply by 100 the unit represents a percentage point change, and if you multiply by 1000 the unit will be 0.1 percentage point change. Which one you choose depends on what kind of change is meaningful in your situation.