A: Taking multiple readings in experiments helps physicists avoid what we call random errors. These are the kinds of errors that are generated in experimental data due to random extraneous sources - for e.g. the surrounding air, slight fluctuations in temperature, delayed human reaction time, etc. Such random errors are what cause data points to be scattered all over if one were to plot them on a graph.
Because random errors are so (you guessed it) random, there isn’t one way to remove them from one particular reading. So if you restrict yourself to just one experimental reading, you have no way of guessing whether a random error has caused it to be greater than or lesser than the true value. To avoid such confusions, we take multiple readings, so that any larger or smaller erroneous values end up balancing out . That gives us a uniform trend-line, or a “line of best fit” that is quite reasonable to use for further processing.