In the companion paper (DOI 10.1021/ma101098d), we presented a quantitative theory for the suppression of the glass transition in a thin polymer film. Our delayed glassification (DG) model follows a proposal by de Gennes that free volume can be transmitted from surface to film interior via kinks transported along molecular strands or loops. In this paper, we use the DG model to predict the effects of molecular weight and film thickness on the film-averaged glass transition for a polystyrene sample. Our predictions for both freestanding and supported films of polystyrene illustrate that the DG model is able to account for some, but not all, of the experimental trends. This leads us to confront a number of issues, including how to average local glass transitions to yield a sample value as well as how to rationalize the nature of the molecular weight dependence for transitions in the thinnest freestanding films.