has parameter β describing the effect of x on the log odds of response in categoryj or below. In this formula, β does not have a j subscript. Therefore, the modelassumes that the effect of x is identical for all J − 1 cumulative logits. When thismodel fits well, it requires a single parameter rather than J − 1 parameters to describethe effect of x.