I believe the graphs are all post-voucher numbers. In other words, those below the red line are the ones who qualified for the vouchers, thus the jumps in ownership. The punchline is that while computer games increased substantially because of the program, using computers for homework did not.