Of course, this assumption here must be established, and the inference method of the whole article will be established. As Wu Songru, a doctoral student in economics at Duke University, pointed out that if the American people in the sample here think that the college budget is also a racial issue that tends to be white, Then the experimental design results here may overestimate the impact of the riots.) Therefore, the experimental group and the control group appear here. The control group is the number of votes in the 1676 district of Los Angeles in 1990 and 1992, and the experimental group is the number of votes in the 1990 and 1992 votes for the public elementary school budget.
When we first whatsapp database calculate the amount of change in the experimental group and the control group within two years, and then compare the "change in the difference" between the two, this gap may come from riots. This research method is called Difference in Difference Design. Illustration of struggle effect Author provided Using this calculation method, the author calculates that, on average, for the 1,676 constituencies in Los Angeles affected by the riots, the percentage of votes in favor of the public and small budget referendums increased by 4.9% more than the percentage of university budget referendums. % (95% confidence interval is 3.7% to 6.0%).
This additional increase, the author believes, is due to violent protests, which has led people to pay more attention to racial issues, so they are willing to vote for a public small budget increase that is more helpful to African Americans. In short, violent protest works, and violent protest reduces racism. Of course, this result may have various explanations, so the author will spend a lot of effort in calculating various possible mechanisms and verifications. First, if violent protests affect public opinion, the more violent the protests, the more changes in public opinion. Indeed, the author estimated the degree of influence of the people according to how close the 1676 constituencies were to the center of violence, and found that the distance was significantly correlated with the amount of change in the impact of votes.