UPDATED 20160505 ca. 8pm

Since I wrote the above post earlier today, it has come to light that Dr. Hersh used the flawed data mentioned in the article below. Even using incorrect data that showed a drop in drunken driving that didn’t exist, Dr. Hersh found no correlation between Uber and Lyft’s presence in Austin and a decrease in drunk driving arrests and accidents. What this means is that Dr. Hersh’s conclusions were correct even though the data set was flawed. The corrected data set, analyzed in the same manner, would result in an even more emphatic absence of the correlation claimed by Uber and Lyft.

I’m guessing that Dr. Hersh’s original data request led to APD’s discovery of the flawed data due to the normal communications back and forth between him and APD personnel.