If you can get rides with Uber instantly versus waiting 20-30mins with the competition, then Uber can raise prices a bit and not lose many customers. People are impatient and Uber has already trained a vast audience that you click a button and get a car in 10mins or less. Once you get used to it, you won't settle for an inferior network where you have to wait 2x or 3x for a couple bucks less.
I think the keyword is significantly. If Uber tripled or quadrupled their prices, I think they'd be fairly vulnerable to regional competitors. Even an increase of 1.5x seems like it'd be hard to pull off.
Speaking anecdotally, I know for sure that I'd use a cheaper alternative to save $5 even if it took an extra 10 min.
It depends on the market you're in. Every market has different economics due to population, density, city layout, socioeconomic landscape, etc.
In many cities Uber is losing money, but in many others they are doing very well. They probably adjust prices to optimize for growth in the cities where they're trying to gain traction. In established cities (as in SF), they tend to raise prices slowly.