A one-tailed test has a less extreme critical value than a two-tailed test, providing a better chance of rejecting the null hypothesis if the observed difference aligns with the predicted direction. However, if the sample mean differs in the unexpected direction, a one-tailed test offers zero chance of rejecting the null hypothesis, unlike a two-tailed test which can detect effects in either direction.