The study of science has been a focal area of concern for educators at most levels in Australia for the past few decades. There has been widespread concern that the number of students studying science at both the secondary and tertiary level is in decline. In particular, concerns have been raised about the low levels of participation in secondary and tertiary level science courses by girls and of the male predominance in subjects such as physics and chemistry. In TIMSS 1995, there were a large number of countries which exhibited significant gender differences in achievement in junior secondary school science, all in favour of boys. In Australia, there were no significant gender differences at this level, which was probably a reflection of the strong campaigns of the 1980s to encourage girls' participation in the sciences. In TIMSS 2003 there were again a large number of countries showing significant gender differences in favour of boys, but a handful in which the differences were in favour of girls. In Australia, however, year 8 boys scored significantly higher than year 8 girls in science. In addition, twice the proportion of boys than girls achieved the advanced international benchmark, and more boys than girls achieved the low benchmark. Girls were outscored by boys in all of the content areas, significantly in all but life sciences (in which internationally, girls significantly outperformed boys), they exhibited lower levels of self-confidence than boys, and did not value science to the same extent as boys. This paper explores gender differences in science achievement of Australian year 8 students; using multilevel modelling to explore the influences on achievement and self-confidence in learning science, separately for males and females.