Specialists in the field of large-scale, high-stakes writing assessment have, over the last forty years alternately discussed the issue of maximizing either reliability or validity in test design. Factors complicating the debate--such as Messick's (1989) expanded definition of validity, and the ethical implications of testing--are explored. An inverse relationship between the loss of reliability and the loss of validity of a test is proffered. The term, Quality, in reference to writing assessment is defined and introduced. Construct complexity is hypothesized as a factor that influences validity, reliability, and quality. It is suggested that the either/or debate concerning emphasis over reliability or validity in test design be put aside in favor of a discussion on how to maximize the quality of an assessment. Insofar as this goal can be achieved, it is necessary in the design of the test to minimize and balance the loss of both validity and reliability. The discussion draws on literature from within the field of writing assessment and from works in the fields of mathematics and information theory. (Contains 6 figures.)