An explanation for the seeming discrepancy could be that the non-tippers take more rides. This would make sense, for example, if there is a population of people

An explanation for the seeming discrepancy could be that the non-tippers take more rides. This would make sense, for example, if there is a population of people who take Uber very frequently (say to commute in a city) and don't usually tip. Perhaps precisely because they use the service so frequently and don't want to pay extra, or because they see the variation in driver quality and want to reserve tips for the best service. Whereas infrequent riders (say to and from the airport) are more likely to tip universally.