A $95\%$ confidence interval for the difference between two population means $\mu_1 - \mu_2$ is $(-2.5, 4.5)$. Based on this interval, what can you conclude about the null hypothesis $H_0: \mu_1 = \mu_2$ at $\alpha = 0.05$?