Critical values of Student’s t distribution by degrees of freedom and significance level. Enter your df and α above and press Look up: the matching cell is highlighted and the exact critical value is shown, even when your df is not one of the rows.
How to read the t table
- Work out your degrees of freedom. For a one-sample or paired t-test, df = n − 1. For an independent-samples t-test with equal variances, df = n₁ + n₂ − 2.
- Decide whether your test is one-tailed or two-tailed and set the Tails option to match. The column headings change to suit.
- Go to the row for your df and the column for your α. That number is the critical value.
- If the absolute value of your t statistic is larger than the critical value, the result is significant at that α.
A worked example
Twelve people try a new diet and the paired t-test on their before and after weights gives t = 2.45. Is that significant at α = 0.05, two-tailed?
- df = 12 − 1 = 11.
- Row 11, column 0.05 gives 2.201. Show this in the table.
- 2.45 is bigger than 2.201, so you reject H₀. The weight change is significant at the 5% level.
Getting a p-value range from the table
A printed table can’t give you an exact p-value, but it can trap it between two columns. Stay on row 11 and move right until you pass 2.45. It sits between 2.201 (α = 0.05) and 2.718 (α = 0.02), so 0.02 < p < 0.05. The exact value is 0.032, which you can get from the p-value from t calculator.
When your df is not in the table
Rows thin out after df = 30. If you have df = 33, the usual rule is to use the next smaller row (df = 30, critical value 2.042). The critical value shrinks as df grows, so using a smaller df gives a slightly stricter test, and you will never claim significance you don’t have. The lookup above gives the exact value for df = 33, which is 2.035.
One-tailed and two-tailed columns
The same number serves two tests. At df = 10, the value 2.228 is the critical value for a two-tailed test at α = 0.05 and for a one-tailed test at α = 0.025. A one-tailed test at α = 0.05 uses 1.812 instead. Only use one-tailed values if you chose the direction before seeing the data.
Why the bottom row looks familiar
The ∞ row holds the z values: 1.645, 1.960, 2.576 and so on. As the sample grows, the t distribution closes in on the normal distribution. With very small samples the difference is large: at df = 2 the two-tailed 0.05 value is 4.303, more than twice 1.96. That extra width is the price of estimating the standard deviation from only a few data points.
Common mistakes
- Using n instead of n − 1 for df. With small samples this changes the answer.
- Comparing a negative t to a positive critical value. For a two-tailed test, compare |t|. t = −2.45 is just as significant as t = 2.45.
- Reading the one-tailed column for a two-tailed test. It halves your α without you noticing and makes results look stronger than they are.
Check yourself
An independent-samples t-test with 15 people per group gives t = 2.10. Significant at α = 0.05, two-tailed?
df = 15 + 15 − 2 = 28. The critical value at row 28 is 2.048. Since 2.10 is larger, yes, it is significant, though only just.
What is the critical value for a one-tailed test at α = 0.01 with df = 20?
Switch Tails to one-tailed and read row 20, column 0.01: 2.528.
A one-sample t-test with n = 8 gives t = 1.90. Roughly what is the two-tailed p-value?
df = 7. On row 7, 1.90 falls between 1.895 (α = 0.10) and 2.365 (α = 0.05), so 0.05 < p < 0.10. It is not significant at 0.05.