One-Tailed Test

Definition(s)

  • In hypothesis testing, we can be interested in a deviation in either direction or in only one direction. If we are interested in either direction (one score is different from another), we use a two-tailed test. If we are interested in only one direction (one score is less than another), and we don’t care if it is greater, then we use a one-tailed test. For example, if we want to know whether a predictive coding system has performed better than chance, then we can use a one-tailed test. We don’t care if the predictive coding system is worse than chance (that would not be particularly useful), only if it is better. Confidence intervals can be one-sided or two-sided as well. The tail refers to the yellow regions in the figure.

    normal-distribution

    A graph of the normal distribution. The confidence interval is in the middle in white. The “tails” are shown in yellow. The 95% confidence interval represents 95% of the area under the curve. In a two-tailed distribution, this 95% area is symmetrically aligned around the average of the distribution. Image from https://en.wikipedia.org/wiki/One-_and_two-tailed_tests  1

Notes

  1. Herb Roitblat, Predictive Coding Glossary.