# Bayesian planned orthogonal contrasts, can one just average two means?

I have a dataset with three levels of one independent variable, repeated measures. To simplify, I have two planned contrasts to perform:

• High Treatment & Low Treatment combine, vs No Treatment
• Hightreatment vs Low Treatment

Or, as contrast coefficients:

HighTreatment(1) + LowTreatment(1) + NoTreatment(-2)
HighTreatment(1) + LowTreatment(-1) + No Treatment(0)


I need a simple way of gaining a Bayes Factor for these planned contrasts. I have been told, that in order to combine the high and low treatment conditions, I am able to simply create a new variable which is the combine mean of High and Low, and then compare this to the no treatment condition (using a Baysian repeated measures t-test).

Thus, my question is twofold:

1. Is the aforementioned method valid? If not,
2. What is an appropriate method to get a Bayes Factor for a specific planned contrasts which weights two variables equally.