I think the paper is more about the "big data problem" and the fact that there was very narrow confidence intervals on data that was off by a lot. So the surve

I think the paper is more about the "big data problem" and the fact that there was very narrow confidence intervals on data that was off by a lot. So the survey technique itself was flawed and shouldn't be relied upon in the future.