Mr the Magnifimoose,

Thanks for the comments. I know the problems or rather the "nonchalance" with which people treat collinearity and the consequent problems in interpretation. However it seems that collinearity in this case was between non-esitmator columns of the X matrix (I think baseline and one movement correction). Which makes me feel better and more willing to gloss over it.

Though one curiosity remains, given the OLS estimation of coefficients, to what extent (if it is) is the mis-estimation of two coeffs (because of collinearity between those two columns of the matrix) of non interest biasing the estimation of the other coefficients (including the ones of interest)?

Cheers

Martin