As rigidity has been regarded as an important variable in the description of personality, its operations and characteristics have been described in detail. It has been frequently described as decreased variability of behavior, e.g., Werner (1946) has emphasized that rigidity is a lack of behavioral variability and that it may be exhibited in varying ways in different situations. Fisher (1950) has observed that flexibility (lack of rigidity) in the individual is . revealed only when he exhibits ability to use alternate modes of response. He further states that rigidity is an overt behavioral manifestation and it is conceived as a single restricted continuum, i.e., all behavior can be classified as the inability to use alternate modes of response. The position taken here is in essential agreement with Fisher's notions. In the present study, rigidity is conceptualized as a reliable and characteristic type of behavior. A person's behavior will be called rigid if he consistently responds in an invariant fashion. Rigidity may be observed in the prejudiced person who reacts in a stereotyped fashion toward individuals in an outgroup, in mental defectives who perseverate on items in a test, in the acquiescent person, and in the negative person who often disagrees with everyone and everything. I t should be noted that the kind of response is not particularly pertinent to the statistical concept of rigidity presently being considered. Rather it is the lack of response variability which is important. The concept of rigidity readily lends itself to statistical definition, e.g., both previously mentioned writers have suggested that rigidity is a lack of variability of response. Specifically, rigidity may be defined as decreased variance of responses in stimulus situations for which no definitive response is apparent to the individual. Rigidity is the failure to make a differential response to differential stimulus patterns. This definition implies that rigidity is nonadaptive behavior. The present definition of rigidity involves two basic assumptions. First, some responses are more similar to the mean or typical response than others. In other words, there are many responses a person may make to a series of stimuli and these responses have varying degrees of similarity to each ocher. For example, a response of like slightly is more similar to like m u c h than is the response dislike slightly. This assumption allows us to apply arith-