I've been asked if I can report the log-ratios of two groups I am comparing in absolute terms to make it easy to understand for the reader. e.g. a log-ratio suggests once group engages in 10% more physical activity, in comparison with the entire study sample.

I fear this is not as simple as calculating 10% of the geometric mean of physical activity for the entire study sample.

Any thoughts?

More Declan Ryan's questions See All
Similar questions and discussions