A recently assembled W7-AS stellarator database has been subject to di
mensionally constrained confinement analysis. The analysis employs Bay
esian inference. Dimensional information is taken from the Conner-Tayl
or (CT) similarity transformation theory, which provides six possible
physical scenarios with associated dimensional conditions. Bayesian th
eory allows the calculation of the probability for each model and it i
s found that the present W7-AS data are most probably described by the
collisionless high-beta case. Probabilities for all models and the as
sociated exponents of a power law scaling function are presented.