Skip to content
Prev 18973 / 20628 Next

Question regarding large data.frame in LMER?

My comment about using poly(Age, 2) instead of I(Age^2) was a
suggestion for trying orthogonal polynomials specifically (i.e., not
raw=T but the default which is raw=F).

The reason is that Age and Age^2 will be very highly correlated,
whereas the linear and non-linear part of poly(Age, 2) are not
correlated.

Can't say if it'll help with your issue, though.
On Mon, 2020-12-14 at 16:00 +0000, Jad Moawad wrote: