‹ BackHN Continuity

Thread

Things that apparently cause cancer

211 points · 78 comments · timpera

  1. aljgz · · focus · HN ↗
    I hoped to see a systematic explanation of what the correlations show. As long as we don't have a viable explanation, there is something interesting there to discover. It can be that the statistical analysis is fundamentally flawed. Pointing out a flaw that tricked several serious researchers is progress. If it was because of some other factor (let's assume there are many more elderly people living around those neighborhoods, can't think of a better example now).
    1. jjk166 · · focus · HN ↗
      > As long as we don't have a viable explanation, there is something interesting there to discover.

      This assumes there is something which needs explanation. For any finite string of gibberish there are an infinite number of languages which can be constructed to ascribe it profound meaning. All patterns in data must be assumed to be illusions until testable predictions made from them are found to be consistently accurate more often than otherwise expected.

Open on Hacker News to reply ↗

Unofficial Hacker News client; not affiliated with Y Combinator.