Table 4.

Difference-in-difference estimate of the effect of yellow on the accident rate by taxi location and lighting condition using Aim(1)Aim(2)=α+βYellowim+εim

Front (1) – Behind (2)Street lighting (1) – Daylight (2)
VariablesParameter estimatesRobust standard errorParameter estimatesRobust standard error
  • The dependent variable is the number of accidents when the taxi was in front (or in street lighting) minus the number of accidents when the taxi was behind (or in daylight) for one driver in a given month, the key independent variable Yellowim is whether the taxi was yellow, α is the constant and εim represents the random error term. The number of observations is based on all registered taxis (ranging from 15,588 to 16,933 per month) in the 36-mo period. R2 is the coefficient of determination. ***P < 0.0001, *P < 0.01.