The purpose of this paper is to use data from the Kagera region of northwestern Tanzania to investigate the long run impact of the timing of parental death on the education outcomes. Apart from defining orphanhood as a continuous variable measured in years a child remains under parental care, we extend the standard specification by (1) breaking down the decision on investing into child's human capital into two stages - a decision to send a child to school, followed by a decision regarding the level of education to be obtained; (2) by allowing the factors affecting the participation decision to be different from those affecting the level decision, at least in the strength of influence. After a brief overview of the key background papers, a description of the mechanisms through which the timing of death can impact the schooling choices of a child, and a description of the data, our analysis begins with a standard double-hurdle model. Our results suggest that there is a statistically significant negative impact of the age at which a child loses his parent on his or her years of schooling, but this impact wears off with time - losing a parent in early childhood is particularly detrimental to schooling, with the six year olds suffering the most; children who lose their parents after the age of 12 experience no negative effects of orphanhood. We also find that wealth, household composition and educational background of both parents have an important effect on orphans' educational outcomes, but that they differ significantly in their impact on the decision to send a child to school vs. the years of completed schooling.