Originally Posted by
OnceAlways
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.
I used NYC as the example. All the other NB fleets show a similar upswing, making the result "statistically significant." It isn't due to random variability. It's company wide, not category specific. NYC 220 was just an easy example.