Comparative results are presented for procedures recently appearing in literature related to standard setting on the National Assessment of Educational Progress--the paper selection method and the contrasting group method. For this comparison, a probabilistic model with normal distribution of performance and a six-point scale were assumed. The paper selection method (American College Testing) requires judges to conceptualize students that are just at the borderline between categories. Papers are selected from a set that represents all levels of performance students at the borderline would likely have produced. The standard is set at the mean of the scores assigned to those papers as part of the regular scoring process. The contrasting groups method requires that teachers internalize the construct to be assessed, and then select students above and below the criterion of success. Papers for those students are scored, and a point between the two score distributions is selected as the standard. For the cases that were modeled, the paper-selection procedure provided better estimates of the standard and the percent above the standard than did the contrasting group procedure. Six tables and 13 figures (graphs) present details of the analyses. (Contains 9 references.) (SLD)