Some theories of human behaviour, particularly economics, tend to assume that humans are fully rational, especially when dealing with subjects such as number, which we suppose to be logical and free of emotion.
Psychological research in recent decades has shown this to be a myth. Many aspects of our behaviour and judgement are driven by heuristics; rules of thumb which are built into our brains. Brains which, remember, for most of human existence were found in the heads of hunter gatherers living in small tribes.
One of the most simple demonstrations of the power of heuristics was established by two of the pioneers in the field, Amos Tversky and Daniel Kahneman. They set up an experiment in which subjects were given a random number between one and one hundred, picked by spinning a wheel, and then asked whether the percentage of African nations in the UN was higher or lower than that number. They were then asked what their estimate was of the actual percentage of African nations in the UN.
They found that, even though the test subjects knew that the first number was random, their second answer was influenced by the first. (In fact, the first number wasn’t random: the wheel was rigged to give only two answers, 10 or 65.)
Further experiments found that this phenomenon applied to other situations, and even worked when the first number was from a completely different context: put on the spot to come up with a number, we will anchor our estimate on a number we heard recently, even if it is entirely irrelevant.