What is a p-value?
The p-value is the probability, if the null hypothesis is true, of getting a result as extreme as yours or more. A small p-value means your result would be unusual if the null hypothesis were true, which is evidence against it. A large one means the data fit the null hypothesis well enough.
Before the test, you choose a significance level, α (often 0.05). If the p-value is less than α, the test rejects the null hypothesis; otherwise it does not.
Source OpenStax Introductory Statistics 2e · 9.4 Rare events, the sample, decision and conclusion ↗
How to calculate a p-value
Here it is with your numbers:
- Area above 2.3163 under the normal curve: 0.010271, which is 0.0103 to 4 decimals
- 0.0103 is less than α = 0.05 at 4 decimals.
It means that if the null hypothesis were true, about 1.03% of samples would give a result at least this extreme.
Which tail? Your alternative hypothesis decides it: “less than” is left-tailed, “more than” is right-tailed, and “not equal to” is two-tailed, with the p-value split evenly between both tails.
Source OpenStax Introductory Statistics 2e · 9.5 Additional information and full hypothesis test examples ↗
Z-score or t-score?
Use a z-score (normal curve) for a test of a mean when you know the population’s standard deviation, and for a test of a proportion. Use a t-score (Student’s t) for a test of a mean when you estimate the standard deviation from your sample; its degrees of freedom are the sample size minus 1.
Source OpenStax Introductory Statistics 2e · 9.3 Probability distribution needed for hypothesis testing ↗
FAQs
What does a 0.05 p-value mean?
If the null hypothesis is true, a result at least this extreme turns up 5% of the time. The smaller the p-value, the stronger the evidence against the null hypothesis. If the p-value is less than your significance level α, reject the null hypothesis; otherwise, don’t.
Source OpenStax Introductory Statistics 2e · 9.4 Rare events, the sample, decision and conclusion ↗
How do I calculate the p-value?
Work out the test statistic, then find the area beyond it in the tail (or tails) your alternative hypothesis points to: left-tailed is the area below, right-tailed the area above, and two-tailed twice the area beyond it on one side. Use the normal curve for a z-score and the t distribution for a t-score.
Source OpenStax Introductory Statistics 2e · 9.5 Additional information and full hypothesis test examples ↗
How to find p-value in Excel with z-score?
For a two-tailed test, =2*(1-NORM.S.DIST(ABS(z),TRUE)). For a right-tailed test, =1-NORM.S.DIST(z,TRUE); for a left-tailed test, =NORM.S.DIST(z,TRUE). NORM.S.DIST(z,TRUE) is the area to the left of z.
How do I calculate the p-value in Excel?
For a t-score, =T.DIST.2T(ABS(t), df) gives the two-tailed p-value and =T.DIST.RT(t, df) the right-tailed one. Microsoft’s example: T.DIST.2T(1.959999998, 60) is 0.054645. For a z-score, use NORM.S.DIST as shown above.
What is p-value vs z-score?
The z-score is the test statistic: it measures how far your sample result is from the value in the null hypothesis, and the p-value turns it into a probability. In OpenStax’s flea shampoo example, z = 2.3163 gives a right-tailed p-value of 0.0103.
Source OpenStax Introductory Statistics 2e · 9.5 Additional information and full hypothesis test examples ↗
Sources
- OpenStax Introductory Statistics 2e: 9.4 Rare events, the sample, decision and conclusion
- OpenStax Introductory Statistics 2e: 9.5 Additional information and full hypothesis test examples
- OpenStax Introductory Statistics 2e: 9.3 Probability distribution needed for hypothesis testing
- Microsoft Support: NORM.S.DIST function
- Microsoft Support: T.DIST.2T function