通常认为由属性和项目关联阵(即Q矩阵)的列对应的项目充任认知诊断测验中行为样本,其实这种做法不能有效防止理想反应模式的误判。如在测验之前便可确定欲测之属性及层级关系,找到可达阵,可证明可达阵的各个列对应的项目类在认知诊断测验中必不可少,否则在理想反应模式下就一定有一些被试会被误判。本文给出充分必要Q矩阵的概念,以区别Tatsuoka(1995,2009)讨论过的充分Q矩阵概念。充分必要Q矩阵才能有效指导测验的编制。 It is generally considered that the item corresponding to the column of attributes and item association matrix (ie, the Q matrix) serves as a sample of behavior in a cognitive diagnostic test. In fact, this practice can not effectively prevent the misjudgment of the ideal reaction pattern. Such as the test before you can determine the properties to be measured and the relationship between levels to find accessibility matrix can prove that the various columns of the corresponding items in the diagnostic test is essential in cognitive tests, or in the ideal response mode Some subjects will be misjudged. In this paper, we give the necessary and sufficient Q matrix concept to distinguish the sufficient Q matrix concept discussed by Tatsuoka (1995, 2009). Fully necessary Q matrix can effectively guide the preparation of the test.