Apologies for being pedantic, but significance is dependent on the distribution and standard deviation of the population being sampled. A swing of 0.4% given the limited total number of trips on NYC220 very likely doesn't meet the actual mathematical threshold of statistical significance.