The present study investigates subjective expectancy of formation of conservative coalitions as a function of status characteristics of actors and nature of the task. Five predictions, derived from Expectation States Theory (cf. Berger, Rosenholz and Zelditch, 1980) are tested. Thirty-four university freshmen participated in the experiment. Subjects estimated the formation of coalitions in eleven ‘choice situations’, which were composed by varying status differences between scenario actors, the nature of the status characteristic and the task to be performed. The predictions derived from Expectation States Theory were generally supported: (1) differentiation on a relevant status characteristic appeared to lead to the expectation of conservative coalitions, (2) a burden of proof process appeared to operate, and (3) the expectation of conservative coalition decreased as the path of relevance was longer. Contrary to predictions, however, it appeared that irrelevant status characteristics did not led to the expectation of conservative coalitions, and that the expectation of conservative coalitions did not covary with the magnitude of the status differentiation. In the discussion of these results two modifications of the burden of proof theorem are proposed and discussed in the light of other findings.