lmBF with polynomial terms

charlesmalpas

Hi All,

I was wondering whether the BayesFactor package for R support polynomial terms in lmBF. For example, I can run:

lm(formula = DV ~ IV1 + poly(IV2, 2, raw = TRUE), data = data)

I can produce a model, but:

lmBF(formula = DV ~ IV1 + poly(IV2, 2, raw = TRUE), data = data)

fails with error "Error in [.data.frame(data, , vars) : undefined columns selected".

I have tried 'manually specifying the the polynomials in the formula (e.g., I(IV2^2)) to no avail.

I have also tried computing the the variables first (e.g., IV22 = IV2^2), which seems to work. I was wondering if I am doing something wrong here with my model specification, or does BayesFactor not yet support polynomials specified using formula.

Many thanks for a fantastic package!

Charles

richarddmorey

This is not currently supported in the formula itself, but you can create the relevant contrasts as columns in your data frame and then include those in the formula.

charlesmalpas

Thanks Richard!