Attribute specification is an important step in making cognitive diagnostic measurement of latent trait. Doing it right will lead to a genuine measurement; otherwise, it may lead to incorrect and reach a wrong conclusion about test takers’ proficiency. This paper introduces the rule space method and applies it to examination degree of attribute misspecification, either exclusion of an essential attribute and inclusion of a superfluous attribute. Simulation studies have shown that either one of the above types of attribute misspecification condition will bear statistical consequences and lead to undesirable specific effects. Particularly this article has demonstrated the importance of assigning a correct order of multiple attributes in order to produce genuine measurement. More investigation are encouraged to understand the mechanism of a biased estimation.