Skip to content

[R-meta] Restricted cubic spline behaviour in moderator variable

1 message · Wolfgang Viechtbauer

#
My preference is to set the knot positions manually anyway. That way, one is always going to use the same positions in all calls of rcs(), rcspline.eval(), and so on. And one can then avoid that knots are set at values like 51.64118 (I would just round this to 52).

Best,
Wolfgang