| Two interrelated problems were investigated in this study. They were (i){A0}the empirical determination of fair and defensible cut scores for the Maryland School Performance Assessment Program (MSPAP), and (ii){A0}an examination of the influence of performance assessments on the teaching-learning process.; Regression analysis and the decision theoretic model were used to estimate optimum cut scores at which the proportions of false-negative and false-positive decisions are minimized. Specifically, 8,204 grade 9 Maryland Functional Reading and Mathematics test scores (MFRT-II and MFMT-II) as well as grade 8 MSPAP test scores were randomly selected. Fifty-percent of the scores (termed screening sample) were used to estimate the regression slopes and constants. The accuracy of the regression equations were tested in the calibration sample. Subsequently, those samples were merged. The combined MSPAP scores were regressed onto the Maryland Functional Tests, then the converse was done. The geometric average line, produced by bisecting the angle between those regression lines, was used to estimate the cut scores. The MSPAP reading cut score of 400 and the classification accuracy of 93.3 were similar to that which were obtained by Hunyh's procedure. This consistency was not observed between the mathematics cut scores; i.e., each method produced a different cut score.; Additionally, grades eight and nine teachers' perceptions of the influence of performance assessment on the teaching process was obtained by using a self-administered questionnaire. Seventy-two public school teachers from a suburban county completed the questionnaire. Between 20--35 percent indicated that MSPAP has so far influenced their selection of topics, the format of their tests and the time spent on the development of critical thinking skills. It is suggested that those positive influences of MSPAP could be enhanced by providing student profiles or feedback on student comprehension deficiencies that may serve as a base for adaptive teaching, curriculum planning and test development. |
