-3

I am using the scam library in R to fit an SCAM model to my data.

If my model contains depth, then I get NA in salinity.

If my model does not contain depth, then I get an estimation for salinity.

Why is this? I don't think it is a convergence issue, since the model converges, but why does the estimation of a variable affect another one?

Improve this question
3
  • 1
    Salinity is usually highly correlated with depth, so you are probably experiencing multicollinearity. But without your data and the R commands that you used, we can only guess Commented yesterday
  • 2
    IMHO, this question is better suited to CrossValidated. Two possible answers to your question: colinearity. Or confounding. Suppose there is a linear relationship between salinity and depth. Then, once you include salinity in your model, adding depth tells you nothing more. And vice versa. You haven't told us if the "vice versa" bit applies, which is one of the many reasons your question cannot be answered definitively (either here or on Cross Validated) as it stands. Commented yesterday
  • 1
    I think we need a reproducible example before we can answer this ... Commented 23 hours ago
Related questions
394
Representing Parametric Survival Model in 'Counting Process' form in JAGS
918
data.table vs dplyr: can one do something well the other can't or does poorly?
405
Convert data.frame columns from factors to characters
Load 7 more related questions Show fewer related questions

0

Reset to default

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.