A series of secondary analyses of data collected during the Second IEA Science Study in 1983/84 were conducted to construct two- and three- level models of the major factors hypothesised to influence the science achievement of Year 12 Australian biology, chemistry and physics students. The models tested were used to compare student achievement in the sciences, and variation in other educational outcomes (eg, self- efficacy, values and attitudes) between: the Australian States and Territories; the Government, Catholic, and Independent education systems; schools located in urban, suburban, non-metropolitan and rural locations; and boys only, girls only and co-educational schools. The responses of male and female students were analysed separately in order to identify some of the sources of gender variation. The secondary issues examined were the influences of student self- efficacy and the effect of sex-type of school. The major factors influencing the science achievement of Year 12 students at level-1 were: computational skills, word knowledge, science self-efficacy, previous science learning experience, attitudes to science, science values and home background. At level- 2 and level-3 the influential factors differed for each of the sexes and each of the sciences. The two-level analyses supported the view that a common 'general' science achievement model operated for each of the sexes. However, the more specific three-level analyses revealed that, while there is some evidence of uniformity at the student-level, the gender differences observed at the class- and school-levels preclude this conclusion. The two-level analyses showed that self-efficacy exerted a positive direct effect on achievement of the same order of magnitude as word knowledge and science attitudes. The three- level results displayed greater variability. In the biology, chemistry and physics models for males self-efficacy was of greater relative importance than the analyses implied. However, self- efficacy factors were dropped during the construction of the three- level model of science achievement for the Year 12 physics and chemistry girls. The sex-type of school attended conferred no comparative advantage with respect to the achievement in science of the Year 12 students.