Since the approval of the Basic Skills Improvement Policy in 1978, the Massachusetts Department of Education has been developing tests and alternative forms for the assessment of student achievement in five basic skills content areas: reading, writing, mathematics, listening, and speaking. Because of the lack of previous research on which to draw in the assessment of speaking skills, the Department previously determined the method by which speaking assessment would be done, and had assessed a large sample of twelfth grade students around Massachusetts using this assessment procedure. This year's study was to investigate reliability and bias issues related to the instrument and to produce final materials for the training of raters to assess speaking. The results demonstrated that there is virtually no bias in the instrument or in the ratings. Raters of all ethnic groups and both sexes exhibited virtually the same pattern of ratings. The recommendation of the study is to finish this work by determining methods to minimize the cost and inconvenience of assessing speaking skills, and to produce a sufficient quantity of materials for use by all districts. Study results relating to rating differences and rater agreement, and statistics describing rater workshop performance in the field are appended. (Author/PN)