Your answer is a good one, you got your three marks for
check results with others
minimise effect of random errors
confirm reproducibility
However, having more results to generate a mean does not in itself make the results more accurate. If you have a zero error making every result biased, that won't go away in the averaging process. Random errors are spread about the mean and so they go away when you do the average.