This study tests the hypothesis of a gender difference in academic achievement as a function of measurement method. The biasing influence of measurement method on achievement has been recognized. Campbell and Fiske (1959) suggested that a considerable proportion of the variation in test scores may be due to features of the form of test (method) used rather than the individual characteristics (traits) which the test is designed to measure. Using a sample of 15-16-year-old Irish school students, gender differences in achievement were examined for three school subjects (Irish, English, and Mathematics), assessed by two methods (multiple choice test, and written public examination). As has been found in other countries, males performed significantly better than females on multiple choice tests, compared with performance on written examinations. An additional hypothesis that the gender difference would be largest for the languages and smallest for mathematics was not supported. This finding runs contrary to an explanation of this phenomenon in terms of greater verbal skills of females. Alternative explanations are proposed and educational policy implications are discussed. (Author/DWH)