The researchers developed a new method called sequential ANOVA to make hypothesis testing more efficient. This method uses a likelihood ratio to gather evidence for both the null and alternative hypotheses, stopping when enough evidence is collected to accept one. Simulations showed that the sequential ANOVA is better than the traditional fixed ANOVA and controls error rates well. Even when assumptions are violated, both methods perform similarly. Overall, the sequential ANOVA is a more efficient way to test hypotheses compared to traditional methods.