She should go through her experimental equipment, setup, and procedure with a fine-tooth comb, to find any little thing that she didn't consider during the first two runs.
Then, she should run the experiment again, one time, two times, or a hundred times more, until she's getting consistent, repeatable results.
