If you're deciding what to test based on your results, you're doing it wrong. Methodology needs to be decided on before the experiment. If you want to change what you're testing, you do it in a separate experiment.