Summary

Imagine reaching into an urn that contains seventy white balls and thirty red\nones, and plucking out ten mystery balls.\n\nPerhaps three of the ten balls will be red, and you’ll correctly guess how many\nred balls total were in the urn. Or perhaps you’ll happen to grab four red\nballs, or some other number. Then you’ll probably get the total number wrong.\n\nThis random error is the cost of incomplete knowledge, and as errors go, it’s\nnot so bad. Your estimates won’t be incorrecton average, and the more you learn,\nthe smaller your error will tend to be.\n\nOn the other hand, suppose that the white balls are heavier, and sink to the\nbottom of the urn. Then your sample may be unrepresentative in a consistent\ndirection.\n\nThat kind of error is called “statistical bias.” When your method of learning\nabout the world is biased, learning more may not help. Acquiring more data can\neven consistently worsen a biased prediction.\n\nIf you’re used to holding knowledge and inquiry in high esteem, this is a scary\nprospect. If we want to...