Analytical strategies for examinining attitudinal data about teacher brinkmanship behavior collected in a semantic-differential format, and delineating the qualitative distinctions among types of brinkmanship behavior are suggested. Teacher brinkmanship is described as behavior which expresses dissatisfaction with facets of the roles teachers are expected to perform; this dissatisfaction is typically manifested so as not to threaten those responsible for administering the situation. Twelve types of behavior are assigned to three categories: (1) subversive or exaggerated obedience; (2) tightroping; and (3) boundary testing. The evaluation of brinkmanship is accomplished by measuring teachers' attitudes toward brinking acts of their colleagues rather than by direct observation of teachers engaged in brinkmanship. Variations of two techniques for two-mode factor analysis, pooled R-technique and pooled T-technique factor analysis, are developed and compared. The pooled R-technique may be submitted to factor-analytic procedures satisfying any of three different objectives: reducing variables to a more manageable set; exploring the feasibility of additional attitudinal parameters; or validating the existence of a conjectured set of attitudinal parameters. When a reasonable amount of certainty exists about the attitudinal dimensions underlying the domain of brinking acts, the pooled T-technique factor analysis can be used. (Author/MH)