# If a t-statistic is negative, should the absolute value be used?

The t-distribution, just like the standard normal, has a mean of $0$. All values to the left of the mean are negative and positive to the right of the mean.