In this study I investigate the birth order effect on educational attainment with careful consideration of the change in schooling differential between men and women over the last century. Using data from the Wisconsin Longitudinal Study, I find that there is a significant negative birth order effect on schooling, and more importantly,the effect is different by gender across the cohorts. Among the cohorts born before 1950, I find that first-borns tend to obtain more schooling than later-borns but that the birth order effect is observed only among sons with no significant difference among daughters especially for small families. For the younger cohorts who were born since the late 1950s, the negative birth order effect is also observed but it is not gender-specific anymore conditional on reversed gender gap in schooling. For the possible source of birth order effect on schooling, other outcomes are also examined. The ability variation between siblings, and aspiration to and preparation for college education in high school years along with parents' expectation about children's schooling are examined and are all estimated to be negatively correlated with birth order.
