Developing scoring rubrics to evaluate student work was studied, concentrating on the use of intermediate points in rating scales. How scales that allow for intermediate points between defined categories should be constructed and used was explored. In the recent National Assessment of Educational Progress (NAEP) visual arts field test, researchers experimented with several formats for constructing scoring rubrics. Some descriptive graphic rating scales (continuous score scales) were pilot tested by 11 raters who scored the NAEP visual arts test for grades 4 and 8. Descriptive graphic ratings were designed to evaluate 4 test production blocks from the assessment, for a total of 50 pieces of student work. The "Facets" computer software was used to analyze the rating data. Raters were able to use the descriptive rating scales reliably. Some of the constructed scales were able to support 7 to 10 rating points rather than the traditional 3 or 4 points. However, there was little appreciable gain in reliability for scales having more than five points. The particular features of the scale (such as defined midpoint) were not as important as the knowledge, skills, and motivation of the rater. An appendix contains the graphic rating scales. (Contains 2 figures, 11 tables, and 32 references.) (SLD)