Standardization by Bivariate Adjustment of Internal Assessments: Sex Bias and Other Statistical Matters

Abstract

The paper canvasses several matters related to the title. In particular, use of scores from the Australian Scholastic Aptitude Test (ASAT) to provide moderating parameters for raw scores derived from teacher assessments of Year 12 students’ achievements is shown to be inconsistent with respect to gender differences of overall achievements when the groups whose scores are moderated have different sex-ratios, as occurs with single-sex male or female or mixed-sex colleges. ASAT scores are examined statistically with respect to gender differences, a rationale for removing the gender bias from the aggregate achievement assessments is described, and the inadequacy of relying solely on aptitude tests to produce comparisons of achievements is noted.