Abstract: A criterion is a sample of performance (including behavior and outcomes), me asured directly or indirectly, perceived to be of value to organizational consti tuencies for facilitating decisions about predictors or programs. Overviewing pr ecursors′ researches about criteria, one conclusion can be drawn that when crit eria are used, we should use multiple criteria. According to this, we put forwar d two hypotheses: (1) Every criterion can’t correlate with all work behaviors. (2) Not only judgemental measurement criteria, but also nonjudgemental measureme nt criteria should be used in performance appraisal. Using critical incident tec hnique, we analysed salesperson’s job in the machinary factory, collected a ser ies of anecdotes of job behavior (collected from supervisors and salespersons) t hat described especially good or especially poor job performance, and sumed up r elative work behavior samples. Organizing these sale behavior samples and other behavior samples, we developed a self-rating forced-choice scale for salesperson ’s performance appraisal. After that, we collected six kinds of judgemental mea surement criteria and ten kinds of nonjudgemental measurement criteria. Judging the correlation between judgemental measurement criteria, nonjudgemental measure ment criteria and behaviors in forced-choice scale, and selecting criteria and w ork behavior samples, we found that all criteria were correlated with at least o ne work behavior sample and those work behavior samples which significantly corr elated with every criterion were not the same as one another. So it verified the hypothesis (1) We also found that only four work behavior samples were correlat ed with judgemental measurement criteria and nonjudgemental measurement criteria . Other work behavior samples either correlated with judgemental measurement criteria or nonjudgemental measurement criteria. So hypo thesis (2) also was verified. Moreover, the factor analysis of judgemental measu rement criteria and nonjudgemental measurement criteria also verified the hypoth esis 2.

Key words: criteria, judgemental measurement, nonjudgemental measurement