Method of Moments---III

• (c) Error in is theoretically

  

  and experimentally

  

  using to be true value of mean and

  to be estimate.

• By central limit theorem, for large N, experimental estimate will give correct error.

• However, if all are the same (always true if ), then experimental error = 0, which is simply wrong.

• The moral is: use theoretical estimate if you only have a few events.