Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Consider reverting back to antilog transformation for m.scale="logit" #80

Open
drnickisaac opened this issue Aug 23, 2021 · 0 comments
Open

Comments

@drnickisaac
Copy link
Contributor

There is an m.scale argument in bma that allows users to record how the data were transformed before going into the model. This argument is then used to determine the back-transformation for the indicator.

Originally, if m.scale="logit" then the inverse logit transformation was used, such that the resulting index would be bounded between 0 and 1 (as the original data going in). However, we changed this to an exponentiation

logit = x <- exp(x), # Counter-intuitively, since we want geometric mean odds

In effect, it means that the indicator values are on the odds scale, not the occupancy scale.

This has caused an external user to misinterpret the results: these are fitted values, so it makes sense for them to be on the same scale as the measurements.

Before making any change we should evaluate the implications. Especially, how are the indicators are currently being run? I think they are using log transformed occupancy data, i.e. m.scale="log", such that the resulting index is the geometric mean occupancy.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

No branches or pull requests

2 participants