In the first of two experiments, subjects screened options designed to distinguish between differential and unit weighting of violations in image theory's compatibility test. In addition, information was presented about the probability that each option would prove to be satisfactory if it were the final choice, framed either as satisfactory or as dissatisfactory. It was found that differential weighting best accounted for the obtained screening data. Probability information also influenced screening, but framing had no effect. The second experiment sought the locus of the influence of the probability information and found that it was treated as an attribute of the options rather than being used to adjust the rejection threshold, a result that is congruent with previous findings.