Question 2. [25 marks] This question is partially based on Duflo (2001, AER), which studies the schooling and labor market consequences of school construction in Indonesia in 1970's, the so-called the INPRES program (1973-1979). Oil crisis in 1970's created large windfall for Indonesia and the government used some of the oil money to fund school construction. About 62,000 schools were built by the government, approximately 1 school built per 500 school-age children. More schools were built in areas which started with fewer schools. Duflo wants to study whether children who were born into areas with more newly built INPRES primary schools get more education, and whether they earn more as adults. Duflo collected data on children born before and after program; Children aged 12 and up in 1974 did not benefit from program while Children aged 6 and under were young enough to be benefited. She found that, for areas with many schools built, the average years of schooling was 8.02 for children over 11 in 1974, and 8.49 for children under 7 in 1974; For areas with few schools built, the average years of schooling was 9.40 for children over 11 in 1974, and 9.76 for children under 7 in 1974. 2.1 [5 marks] Version A. What are the control and treatment groups in this empirical application? Explain briefly. 2.2 [5 marks] Version B. In order for the difference in differences (DID) method to work, the conditional mean independence assumption must be satisfied. Briefly explain what this assumption means in this application. Question 2. [25 marks] This question is partially based on Duflo (2001, AER), which studies the schooling and labor market consequences of school construction in Indonesia in 1970's, the so-called the INPRES program (1973-1979). Oil crisis in 1970's created large windfall for Indonesia and the government used some of the oil money to fund school construction. About 62,000 schools were built by the government, approximately 1 school built per 500 school-age children. More schools were built in areas which started with fewer schools. Duflo wants to study whether children who were born into areas with more newly built INPRES primary schools get more education, and whether they earn more as adults. Duflo collected data on children born before and after program; Children aged 12 and up in 1974 did not benefit from program while Children aged 6 and under were young enough to be benefited. She found that, for areas with many schools built, the average years of schooling was 8.02 for children over 11 in 1974, and 8.49 for children under 7 in 1974; For areas with few schools built, the average years of schooling was 9.40 for children over 11 in 1974, and 9.76 for children under 7 in 1974. 2.1 [5 marks] Version A. What are the control and treatment groups in this empirical application? Explain briefly. 2.2 [5 marks] Version B. In order for the difference in differences (DID) method to work, the conditional mean independence assumption must be satisfied. Briefly explain what this assumption means in this application