Consider a study comparing two group means with independent samples. If the pooled variance $s_p^2$ is used, what assumption is being made about the populations?