Statistical table
T-Table (Student’s t Distribution Table)
T table of one- and two-tailed critical values for df 1–1000 and ∞, with confidence levels, an exact lookup and p-value, worked examples and a free PDF.
Updated
Quick answer
A t table lists critical values of Student’s t distribution by degrees of freedom (rows) and significance level α (columns). Find the row for your df, choose the one-tailed or two-tailed α column, and read the value where they meet. For example, a two-tailed test at α = 0.05 with df = 15 has critical values ±2.131. Reject H₀ when your t statistic falls beyond the critical value.
Find any t critical value
Decimals allowed (e.g. Welch df).
Shows the exact p-value and the decision.
Two-tailed t critical value, df = 15, α = 0.05
±2.1314
Confidence level 95%.
What is a t-table?
Student’s t distribution is the sampling distribution that applies when a mean is standardized with the sample standard deviation s instead of the population σ. It is bell-shaped and symmetric like the normal curve but has heavier tails, and its exact shape depends on the degrees of freedom (df). Small samples have few degrees of freedom and fatter tails, so they need larger critical values to reach the same significance level. The same table is known as the t distribution table, the t-score table, the t chart, the table of t statistics, the t critical value table or, in AP Statistics, the t-star table (t* is the critical value), and it is sometimes searched for simply as "ttable".
Each cell of this table is the critical value t_c with an upper-tail area of α, so P(T ≥ t_c) = α for a one-tailed test. Rows cover df 1 to 30, then 40, 50, 60, 80, 100, 120, 1000 and ∞. Columns cover one-tailed α = 0.25, 0.10, 0.05, 0.025, 0.01, 0.005, 0.001 and 0.0005; the two-tailed view labels the same columns 0.50 to 0.001 and shows the matching confidence level, from 50% to 99.9%. For df values that are not listed, use the exact lookup.
How to read the t-table
- Calculate the degrees of freedom for your procedure, for example df = n − 1 for a one-sample or paired t-test.
- Decide whether the test is one-tailed or two-tailed, and switch the table to the matching view.
- Find the column for your significance level α. For a confidence interval, use the two-tailed column with α = 1 − confidence level, for example 0.05 for 95%.
- Read across the df row to that column to get the critical value t_c. For a two-tailed test, the rejection region is t ≤ −t_c or t ≥ t_c.
- If your df is not listed, use the next smaller df in the table, which is slightly conservative, or the exact lookup.
Examples
Two-tailed test
Problem. A one-sample t-test uses n = 16 observations at α = 0.05, two-tailed. What is the critical value?
- 01df = n − 1 = 15.
- 02Two-tailed α = 0.05 is the same column as one-tailed α = 0.025.
- 03Row 15 in that column reads 2.131.
Answer. Reject H₀ if t ≤ −2.131 or t ≥ 2.131.
One-tailed test
Problem. A pooled two-sample t-test has n₁ = 12 and n₂ = 14, H₁: μ₁ > μ₂ and α = 0.05. Find the critical value.
- 01df = n₁ + n₂ − 2 = 24.
- 02Use the one-tailed α = 0.05 column, row 24: 1.711.
Answer. Reject H₀ if t ≥ 1.711.
Confidence interval multiplier
Problem. Find the t multiplier for a 95% confidence interval for a mean based on n = 20 observations.
- 01df = 20 − 1 = 19.
- 02A 95% interval leaves 2.5% in each tail, so use two-tailed α = 0.05 (one-tailed 0.025).
- 03Row 19 reads 2.093.
Answer. The interval is x̄ ± 2.093 · s / √20.
One-tailed vs two-tailed t table and confidence levels
A one-tailed (one-sided) test puts all of α in one tail. A two-tailed (two-sided) test splits α between both tails, so its critical value equals the one-tailed critical value at α/2. That is why one set of numbers serves both views, with two sets of column labels. The same two-tailed column gives the t multiplier for a confidence interval: 95% confidence uses two-tailed α = 0.05. The bottom row, df = ∞, matches the standard normal critical values, which shows how the t distribution approaches the normal as df grows.
| One-tailed α | Two-tailed α | Confidence level | t_c at df = 10 | t_c at df = ∞ |
|---|---|---|---|---|
| 0.25 | 0.50 | 50% | 0.700 | 0.674 |
| 0.10 | 0.20 | 80% | 1.372 | 1.282 |
| 0.05 | 0.10 | 90% | 1.812 | 1.645 |
| 0.025 | 0.05 | 95% | 2.228 | 1.960 |
| 0.01 | 0.02 | 98% | 2.764 | 2.326 |
| 0.005 | 0.01 | 99% | 3.169 | 2.576 |
| 0.001 | 0.002 | 99.8% | 4.144 | 3.090 |
| 0.0005 | 0.001 | 99.9% | 4.587 | 3.291 |
Degrees of freedom for common t procedures
Welch’s test usually produces a fractional df. Either round it down to use the table or use the exact lookup, which accepts fractional values. The Degrees of Freedom Calculator works out df for each procedure below.
| Procedure | Degrees of freedom |
|---|---|
| One-sample t-test or t interval | n − 1 |
| Paired t-test | n − 1, where n is the number of pairs |
| Pooled two-sample t-test | n₁ + n₂ − 2 |
| Welch’s t-test | Welch–Satterthwaite df, usually not a whole number |
| Slope in simple linear regression | n − 2 |
| Test of a Pearson correlation | n − 2 |
When your df is not in the table
Above df = 30 the table lists only selected rows. Using the next smaller listed df gives a slightly larger critical value, so the test becomes a little conservative. For df = 45 at two-tailed α = 0.05, the df = 40 row gives 2.021, while the exact value is 2.014. For df = 49 at one-tailed α = 0.005, the df = 40 row gives 2.704, while the exact value is 2.680.
For an exact critical value at any df or α, use the lookup on this page or the T Critical Value Calculator. To go the other way, from a t statistic to a p-value, use the P-Value from t Calculator.
Finding a p-value with the t table
A t table gives critical values, not p-values, but it can bracket the p-value. Go to the row for your df, find the two columns whose values bracket |t|, and read their α labels: the p-value lies between them. Use the two-tailed labels for a two-tailed test.
With df = 15 and t = 2.5, the row reads 2.131 at two-tailed α = 0.05 and 2.602 at two-tailed α = 0.02, so 0.02 < p < 0.05. The exact two-tailed p-value is 0.0245. Enter your t statistic in the lookup above to get the exact p-value and the decision.
t table vs z table vs chi-square table
All three are tables of critical values, but each belongs to a different test statistic. Pick the table that matches the statistic your test produces.
| Table | Statistic | Use it when |
|---|---|---|
| t table | t | Testing or estimating a mean (or a difference of means, a regression slope or a correlation) with σ estimated from the sample |
| z-table | z | σ is known, testing a proportion with a large sample, or df is large enough that t ≈ z |
| Chi-square table | χ² | Goodness-of-fit and independence tests on counts, or a test or interval for one variance |
Common errors when reading a t table
- Using the one-tailed column for a two-tailed test. At df = 15 and α = 0.05 this gives 1.753 instead of 2.131 and doubles the real Type I error rate.
- Using n instead of df. For a one-sample test with n = 16, the row is 15, not 16.
- Rounding df up. If df = 45 is not listed, use the df = 40 row (or the exact value), not df = 50, which gives a critical value that is too small.
- Using z when the sample is small. With df = 10, the 95% multiplier is 2.228, not 1.960; the z value gives an interval that is too narrow.
- Comparing a negative t with a positive critical value. For a two-tailed test compare |t|; for a left-tailed test compare t with −t_c.
t table practice problems
Try each one with the table, then check your answer with the lookup.
- A paired t-test has 9 pairs, two-tailed α = 0.05. Critical value? Answer: df = 8, ±2.306.
- What is the t multiplier for a 99% confidence interval with n = 25? Answer: df = 24, two-tailed α = 0.01, 2.797.
- A left-tailed one-sample test has n = 13 and α = 0.01. Critical value? Answer: df = 12, reject H₀ if t ≤ −2.681.
- A right-tailed test gives t = 1.90 with df = 22. Is it significant at α = 0.05? Answer: yes, 1.90 > 1.717; the row brackets p between 0.025 and 0.05 (exact p = 0.0353).
Print or download the t table
Use Download PDF at the top of the page for a printable t table: it opens Table 2 of the free Statculator statistical tables PDF, with the same 40 rows and both one-tailed and two-tailed α labels. Print / save as PDF prints this page, including the view you have selected.
Every value is computed from the exact Student t inverse CDF and rounded to three decimals, so it can differ from older printed tables in the last digit. You can check any entry against the NIST/SEMATECH e-Handbook table of critical values of the t distribution; its t distribution page gives the formulas. In software, Excel’s T.INV.2T(α, df) and R’s qt(1 − α/2, df) return the two-tailed value.
Frequently asked questions
How do I find the degrees of freedom for a t table?
What is the t critical value for a 95% confidence interval?
Why are t critical values larger than z critical values?
What if my t statistic is negative?
Can I get a p-value from a t table?
What is t* in a t table?
Cite this page
Statculator. (2026, November 14). T-Table (Student’s t Distribution Table). Retrieved November 14, 2026, from https://statculator.com/tables/t-table/
Skip the table: exact calculators
- T Critical Value Calculatort* for any degrees of freedom and α, beyond what printed tables show.
- P-Value from t Calculatorp-value for a t statistic and its degrees of freedom.
- T-Test CalculatorOne-sample, two-sample (pooled or Welch) and paired t-tests in one place.
- One-Sample t-Test CalculatorTest a mean against a hypothesized value when σ is unknown.
Values are computed from the exact distribution functions used by Statculator’s calculators (see methodology).