It depends on the school of economics. The Austrian school makes no such assumption which is why it lacks specific predictive ability. It is based on how humans act which is basically they pick from what they perceive as the best choice of those they see as available at the time. What this doesn't allow is your judgement as to what is rational. If someone decides at a certain time smoking a bowl is preferable to going to work at that moment than that is what they do otherwise they would do something else. The problem with this is all choices are ordinal. Basically you can rank the choices based on what they did but you can't compare them to a non choice.
For example if someone went to the store and bought 4 apples for $2 all you can say is they preferred the apples to the $4 and time end effort to go to the store. You can draw no other conclusions.
But most economists hate this because it eliminates specific predictive power. For example you can predict that if the Fed creates a $1T and gives it to the banks there will be a bubble somewhere and a misallocation of resources into that bubble. But you have no clue where the bubble will be or when it will pop. The only conclusion you can reach is not to artificially create money in the first place and let the markets set interest rates and money supply.