VerifiedEvidence: highv1.0.0

One-Tailed Test

A hypothesis test assessing evidence for an effect in only one specified direction, requiring strong prior justification for that direction.

Last reviewedDarrin Baines IP Ltd

Concept Architecture

Concept


Theoretically, a One-Tailed Test is a statistical hypothesis test in which the alternative hypothesis specifies a directional effect, restricting statistical evidence to one tail of the sampling distribution. The test evaluates whether a population parameter is either greater than or less than a specified value, but not both. One-tailed tests exist to increase statistical power for detecting effects in a prespecified direction when departures in the opposite direction are considered irrelevant to the research question.

Mathematically, a one-tailed test is defined by a null hypothesis and a directional alternative hypothesis. The entire significance level is allocated to a single tail of the sampling distribution, and the decision to reject the null hypothesis is based on whether the observed test statistic falls within that critical region. Test statistics such as the z statistic, t statistic, Wald statistic or likelihood ratio statistic are evaluated against the corresponding one-sided reference distribution.

In practice, one-tailed tests are used only when the research hypothesis specifies a justified directional expectation before data analysis. In health economics, they may be applied when evaluating whether a new intervention reduces healthcare costs, improves health outcomes or increases cost-effectiveness relative to a comparator, provided that effects in the opposite direction would not alter the study objective. Regulatory and methodological guidance generally recommends two-tailed tests unless there is strong justification for a directional hypothesis.

Purpose


Used to evaluate directional research hypotheses, increase statistical power for detecting effects in a specified direction and perform one-sided statistical inference in health economic and clinical research.


Mathematical Formulae

Primary Formula

Upper-tailed hypothesis:

H?: ? � ??

H?: ? > ??

Lower-tailed hypothesis:

H?: ? � ??

H?: ? < ??

Generic test statistic:

Z = (?? ? ??) / SE(??)

Supporting Formulae

Upper-tail p-value:

p = P(Z � zobs)

Lower-tail p-value:

p = P(Z � zobs)

Decision rule:

Reject H? if p � �

Related Mathematical Methods

  • Two-Tailed Test
  • Hypothesis Testing
  • z-Test
  • t-Test
  • Wald Test
  • Likelihood Ratio Test
  • Statistical Power

Example

A health economist evaluates whether a new disease management programme reduces annual healthcare costs compared with standard care.

The hypotheses are:

H?: ?Treatment � ?Control

H?: ?Treatment < ?Control

The observed test statistic is:

t = ?1.92

The corresponding one-tailed p-value is:

p = 0.028

Using a significance level of � = 0.05, the null hypothesis is rejected. The analysis provides evidence that the programme reduces annual healthcare costs relative to standard care in the prespecified direction.


Excel Implementation

FunctionExample FormulaHealth Economics Application
T.TEST=T.TEST(B2:B101,C2:C101,1,3)Perform a one-tailed t-test comparing mean healthcare costs or outcomes.
NORM.S.DIST=1-NORM.S.DIST(D2,TRUE)Calculate an upper-tail p-value from a z statistic.
T.DIST.RT=T.DIST.RT(D2,98)Calculate a one-sided p-value from a t statistic.
IF=IF(E2<=0.05,"Reject H0","Do not reject H0")Apply the decision rule for a one-tailed hypothesis test.
ABS=ABS(D2)Summarise the magnitude of the test statistic before directional interpretation.

VBA (Optional)

A VBA routine can automate one-tailed hypothesis testing across multiple health economic outcomes and produce summaries of test statistics, p-values and statistical decisions.


Sources

  • Casella G, Berger RL. Statistical Inference. Cengage Learning.
  • Lehmann EL, Romano JP. Testing Statistical Hypotheses. Springer.
  • Altman DG. Practical Statistics for Medical Research. Chapman & Hall.
  • Fisher RA. Statistical Methods for Research Workers. Oliver and Boyd.
  • Briggs A, Claxton K, Sculpher M. Decision Modelling for Health Economic Evaluation. Oxford University Press.
  • Drummond MF, et al. Methods for the Economic Evaluation of Health Care Programmes. Oxford University Press.

Library

Publications

1
  • Book

    Statistical Analysis of Cost-Effectiveness Data — Willan & Briggs, 1st Edition ed., 2006 (John Wiley & Sons)

    A synthesis of statistical methods for analysing cost-effectiveness data, including net-benefit regression, confidence intervals for the ICER, cost-effectiveness acceptability curves, and covariate adjustment. Part of the Wiley Statistics in Practice series.

Frequently Asked Questions (6)

  • What is a one-tailed test?

    A hypothesis test assessing evidence for an effect in only one specified direction, requiring strong prior justification for that direction.

    Source: Neyman & Pearson 1933

  • What does a one-tailed test look for in only one direction?

    A one-tailed test looks for evidence of an effect in a single, prespecified direction, for instance that a treatment improves an outcome, ignoring the possibility that it makes things worse. Concentrating the test's attention on one side makes it easier to reach significance for an effect in that direction, but only if the direction was justified in advance. It is used cautiously because choosing one tail after seeing the data, or to scrape past a threshold, inflates the false-positive risk. Testing in one direction only is its nature. Kirkwood and Sterne (2003) describe this.

    Source: Kirkwood & Sterne 2003

  • When is a one-tailed test appropriate?

    A one-tailed test is appropriate only when there is strong prior justification for considering an effect in a single direction and an effect in the opposite direction would be irrelevant or implausible, so that testing only one side is warranted. So a one-tailed test is appropriate in the limited circumstances where the direction of any effect is confidently predicted and the reverse would not matter, which is uncommon, since usually an effect in either direction is of interest, and this is why one-tailed tests are used cautiously and require justification, as choosing one merely to gain significance is inappropriate.

    Source: Neyman & Pearson 1933

  • How does a one-tailed test differ from a two-tailed test?

    A one-tailed test assesses evidence for an effect in only one specified direction, placing the whole significance level in one tail, while a two-tailed test assesses evidence for an effect in either direction, splitting the significance level between both tails. The one-tailed test has more power to detect an effect in its chosen direction but cannot detect one in the other. So the two differ in whether they consider one or both directions, with the one-tailed test more powerful in its direction but blind to the opposite, and the two-tailed test more conservative and generally preferred unless a one-sided alternative is strongly justified.

    Source: Neyman & Pearson 1933

  • Why are one-tailed tests used cautiously?

    One-tailed tests are used cautiously because they cannot detect an effect in the unconsidered direction and because choosing a one-tailed test can make it easier to reach significance, so using one without genuine prior justification, or switching to it after seeing the data, is a misuse that inflates the false positive risk. So one-tailed tests are treated with caution, since their added power in one direction comes at the cost of ignoring the other, and their potential for abuse means they require a prespecified, well-justified reason for testing only one side, which is why two-tailed tests are the default in most analyses.

    Source: Neyman & Pearson 1933

  • What are the risks of using a one-tailed test?

    The risks of using a one-tailed test include missing a real and important effect in the direction not tested, since the test cannot detect it; and, if the choice is not properly justified and prespecified, inflating the chance of a false positive by effectively lowering the threshold in the chosen direction. So the risks of a one-tailed test are overlooking effects the other way and misusing the approach to obtain significance, which is why it is applied only with strong, prespecified justification, and why two-tailed tests, which consider both directions and are more conservative, are generally preferred to guard against these risks.

    Source: Neyman & Pearson 1933

Trust Record

Verified by Dr Darrin Baines

British health economist

Professional identity: darrinbaines.org

Verification date: 19 Dec 2025

Content version: 1.0.0

Canonical Identity

Term code
HE-ES-SA-142

Stable URI · Machine-readable · Resolvable · CC BY 4.0