A researcher conducts an experiment to determine if a new active-recall study method improves test scores more than passive reading. The experimental group uses the active-recall method and scores an average of 87% on a test. The control group uses passive reading and scores an average of 84%. Based solely on these averages, the researcher concludes that the active-recall method is definitively more effective. What is the primary flaw in the researcher's reasoning, and what must be done before a valid conclusion can be drawn?