A study examined whether composition specialists can counterbalance the potential privileging of the assessment perspective, or of self-appointed interpreters of that perspective, through the study of assessment discourse as text. Fourteen assessment texts were examined, most of them journal articles and most of them featuring the common methodology of educational testers--statistical inference. Critical strategies employed were: closely examining the language in which assumptions about writing was couched; considering the scholarly authority used to support an assumption; juxtaposing an assumption made by an assessor with a representative text from the composition discourse community; analyzing the "works cited" list of a discourse; and habitually seeking definitions for those qualities or abilities assessors have set out to measure. Conclusions, admittedly impressionistic, include: (1) it is most fruitful to critically examine the assumptions about writing which inform the tests in question; (2) three of the assessment texts had no citations familiar to someone who has been reading composition literature for almost two decades; and (3) many assessors are less definite and assertive about the process of evaluating writing than composition specialists have feared. Writing teachers could be handicapped by a commitment to individual perception and a reputation for idiosyncratic grading unless composition specialists can help their audiences see the value of those traits, through an informed criticism based on a knowledge of composition theory and a familiarity with assessment texts. (Twenty references are attached.) (RS)