When a theory that has a huge amount of evidence for it, if you find a data point that contradicts it, the most likely hypothesis is sometimes still that the theory is correct. The question becomes “what would make both the data point and the theory correct?” (whether in the environment or your instrumentation) That’s not confirmation bias, that’s generating the most likely hypotheses to investigate them first.
For example, when Uranus was discovered, it didn’t follow what Newtonian gravity predicted. Scientists asked themselves “what other mass would need to be where to explain Uranus’ orbit under Newtonian gravity” and that allowed them to discover Neptune.
Same happened with Mercury. So scientists went on to search for planet Vulcan, but never found it. (Although some people actually reported observing it at the time!) In that case, Newtonian gravity was actually failing to explain the phenomenon, and it required general relativity to fully explain Mercury’s orbit because of the curvature of spacetime near the Sun had measurable effects.
Of course, Newtonian gravity still has a lot of explanatory power—all the previous predictions it made were still true, but it was further generalized in new environments that were previously outside the distribution of experiments.
When investigating an unexplained data point in a well proven theory, consider asking yourself “Are we looking for Neptune or Vulcan?”
Are you looking for Neptune or Vulcan?
When a theory that has a huge amount of evidence for it, if you find a data point that contradicts it, the most likely hypothesis is sometimes still that the theory is correct. The question becomes “what would make both the data point and the theory correct?” (whether in the environment or your instrumentation) That’s not confirmation bias, that’s generating the most likely hypotheses to investigate them first.
For example, when Uranus was discovered, it didn’t follow what Newtonian gravity predicted. Scientists asked themselves “what other mass would need to be where to explain Uranus’ orbit under Newtonian gravity” and that allowed them to discover Neptune.
Same happened with Mercury. So scientists went on to search for planet Vulcan, but never found it. (Although some people actually reported observing it at the time!) In that case, Newtonian gravity was actually failing to explain the phenomenon, and it required general relativity to fully explain Mercury’s orbit because of the curvature of spacetime near the Sun had measurable effects.
Of course, Newtonian gravity still has a lot of explanatory power—all the previous predictions it made were still true, but it was further generalized in new environments that were previously outside the distribution of experiments.
When investigating an unexplained data point in a well proven theory, consider asking yourself “Are we looking for Neptune or Vulcan?”
Related “Why Science Doesn’t Make Laws Anymore”: https://www.youtube.com/watch?v=EVJdwD7coQ4