In response to growing concern about the lack of basic writing skills, this paper presents an overview of the issues involved in selecting a method for the assessment of students' writing skills. After general criteria for determining the appropriateness of a writing evaluation procedure are outlined, the merits and limitations of objective tests and direct evaluation of writing samples are presented. While acknowledging the speed and ease with which objective tests can be administered, the paper emphasizes that only a writing sample permits a full assessment of writing ability and warns of the lack of correlation between direct and indirect measures. Next, a review is presented of the advantages and disadvantages of three scoring options for writing samples: (1) analytic scoring, which judges papers in terms of ideas, mechanics, organization, wording, and flavor; (2) holistic scoring, which responds to the writing sample as a whole; and (3) primary-trait scoring, a form of holistic scoring which considers the writing sample as a particular type of discourse and rates it accordingly. Finally, the paper discusses the political aspects of beginning a testing program, pointing out the need for faculty support of the program, for their willingness to participate as readers, and for consensus on the choice of a scoring option. Concluding comments argue that while no method of measurement is perfect, faculty have a responsibility to educate themselves and the administration about the scoring options available and to influence their institutions to utilize direct measures of writing ability. (AYC)