When you have a confidence interval width equal to one minus the significance level, if the hypothesized population parameter is within the confidence interval, you should fail to reject the null hypothesis.