This paper investigates the effect of sibling composition and gender wage differentials on the higher education of women in Japan using a micro data set, the 1998 and 2003 "National Survey on Families ". In Japan, the female enrolment rate at four-year universities is 47.5%, much lower than its male counterpart of 66.5%. So, parents possibly take different educational strategies between their sons and daughters. We focus on female higher education because 96% of students in Japan enroll high school, so it is important which way parents choose after their children's graduation of high school. There are two types of school in higher education in Japan, one is two-year college, junior college and the other is four-year college, university. University graduates are likely to earn more than junior college graduates, so we analyze junior college and university separately.In Japan, compulsory education is under the 6-3 school system (six-years elementary school and three-years junior high school), and is up to age 15. From 1975, the enrolment rate at high school has been about 96%, so most students continue studying beyond the compulsory level. Parents may decide whether they have their children entered the labor market or attended school after their children's graduation of high school. In detail, there are three options; entering the labor market, attending a junior college or vocational college, or attending a university.Attendance at both university and junior college are included in higher education, but they are very different. Junior college means two-year course, and university means four-year course, and both junior college and university are regarded as higher education. However, it should note that university graduates rather than junior college graduates are provided access to "career" job. Because junior colleges offer only non-vocational subjects and universities offer broader curricula and vocational subjects. Thus, most firms recruit junior college graduates and university graduates separately and university graduates are likely to earn more than junior college graduates (Edwards et al., 1993).This paper makes three contributions. First, using detailed information of sibling composition, we investigate the effect of sibling composition on educational achievement in Japan. Second, gender wage differentials are considered. In theoretical models discussed earlier, when making their allocations parents consider the expected rate of return or lifetime income of their children. Thus, differences in the expected wages between sons and daughters may influence the resources allocated to them by their parents. Third, we use a more recent birth cohort to examine rapid increase of the female enrolment rate of university.The results show that: (1) being female increases the probability of junior college enrolment, while decreases the probability of university enrolment. Parents may regard junior college and university as different things and take different strategies according to sex of their children. However, having brothers have no effect on female educational achievement. Parents might decide their resource allocation according to sex of their children, but do not care sibling composition; (2) the wage differentials variables have no relationship with the probability of junior college enrolment or university enrolment. Only exception is wage differentials of aged 50-54, and wage differentials of aged 50-54 decrease the probability of university of both sexes; (3) the income and educational level of parents have important roles in influencing the educational levels. The high income of fathers has positive impact on the years of educational level and the possibility of university attendance. The educational levels of parents are also important determinant of attending a junior college or a university. Having a mother who was a university graduate decreases the probability of junior college enrolment, while having a father who was a university graduate increases the probability of junior college enrolment. Having parents who were high school graduates and university graduates increase the probability of university enrolment.