Free t-value calculator: get the critical t-value, p-value, and t-test results (t-stat, CI, Cohen’s d) instantly from your sample or degrees of freedom.
Student’s t‑Value Calculator
Exact critical values, p‑values and t‑tests. Computed live, never read off a printed table.
Why you can trust these results: this calculator does not read numbers from a printed t table or guess between rows. It solves the Student’s t distribution directly for whatever degrees of freedom you enter. We checked it against SciPy and mpmath on more than 300 test cases, and it reproduces the worked examples in the NIST Engineering Statistics Handbook (for example T = −12.62059 with 326 degrees of freedom). Press Show working in the calculator to see every step for yourself.
Table of contents
- T value calculator: what it does
- t-value, t critical value and p-value
- The t distribution and degrees of freedom
- How to use this calculator
- How to find the t value in statistics
- One-tailed vs two-tailed critical values
- t table vs z table
- t distribution table
- Confidence interval example
- Worked t-test examples
- Welch, pooled or one-sample?
- Cohen’s d
- Common mistakes
- Related tools
- Frequently asked questions
- Add this tool to your website
T value calculator: what it does and who it is for
Anyone who has sat a statistics module remembers the t table at the back of the textbook. You pick a row for the degrees of freedom, slide along to the right column, squint, and hope you did not land one line off. A t value calculator takes that step away. The calculator above is a free Student’s t value calculator that gives the exact answer for any degrees of freedom, and then goes a few steps further.
It works as a t critical value calculator, as a p value calculator for any t score, and as a t test calculator that handles one-sample, Welch and pooled two-sample tests. You type in a few numbers, the answer appears straight away, and a live picture of the t distribution shows what the number means. Think of it as a free online t table calculator: nothing to install, no account to create, and no need to flip to the back of a book.
Here is what you can do with it:
- Find the critical t value for a confidence level, one-tailed or two-tailed, and see the working behind it (a t critical value calculator with steps).
- Turn a t statistic and its degrees of freedom into an exact p-value, which is the quickest way to get a p value from a t score.
- Run a one sample t test, a Welch t test or a pooled (equal variance) two sample t test from means, standard deviations and sample sizes.
- Get a confidence interval and Cohen’s d next to every t-test result.
- Copy the result, or a ready-made APA-style line, with one tap.
t-value, t critical value and p-value: three numbers that get mixed up
Most of the confusion around t tests comes from three similar-sounding numbers. They answer different questions, so it is worth keeping them apart.
| Name | Also called | What it tells you |
|---|---|---|
| t statistic | t value, t score | How many standard errors your sample result sits from the value you are testing |
| Critical t value | t critical, t* | The cut-off your t statistic has to pass before the result counts as significant |
| p-value | probability value | How surprising your data would be if the null hypothesis were true |
| Degrees of freedom | df | Sets the exact shape of the t curve you are using |
The t statistic itself is simple. You take the gap between your sample mean and the value you are testing against, then divide by the standard error:
Here s is the sample standard deviation and n is the sample size. If you were looking for a t statistic calculator or a t score calculator, this is the number you want, and the t-test tab in the calculator works it out from nothing more than a mean, a standard deviation and a sample size. That also makes it a quick t value calculator from mean and standard deviation.
People often ask about t score vs z score. A z score assumes you know the true population standard deviation. In real projects you almost never do, so you use the sample standard deviation instead, and that extra uncertainty is exactly what the t distribution accounts for. With a very large sample the two give nearly the same answer.
The name comes from William Sealy Gosset, who worked at the Guinness brewery in Dublin and published the method in 1908 under the pen name “Student”. He needed a reliable way to judge small samples of barley and beer, which is why the t distribution is built for small-sample work.
The t distribution and degrees of freedom
The t distribution looks like the normal bell curve, but with heavier tails. Heavier tails mean extreme values are more likely, so you need a bigger t statistic before you can call a result unusual. How heavy the tails are depends on the degrees of freedom. With very few degrees of freedom the curve is flat and wide. As the number grows, the curve tightens and slowly becomes the normal curve.
Curves for 2, 5 and 30 degrees of freedom. The dashed line is the standard normal curve, which the t curve approaches as df grows.
If you need a degrees of freedom calculator, the rule depends on the test you are running:
| Situation | Degrees of freedom |
|---|---|
| One sample, or a paired test on differences | n − 1 |
| Confidence interval for a mean | n − 1 |
| Two samples, equal variances (pooled) | n1 + n2 − 2 |
| Two samples, unequal variances (Welch) | Welch–Satterthwaite formula, usually a decimal |
You do not have to do this arithmetic yourself. In the Critical t and p-value tabs, the from n and from n1, n2 buttons turn sample sizes into degrees of freedom for you. In the t-test tab it is calculated automatically.
How to use this t value calculator, step by step
At the top of the calculator you will see three tabs: Critical t, p-value and t-test. Each tab answers a different question, and the dark panel on the right always shows the answer together with a curve. A small Reset button next to the tabs puts everything back to the starting values.
Mode 1: Critical t (the t critical value calculator)
Use this tab when you already know your confidence level or significance level and want the cut-off value, for example when you are building a confidence interval or setting up a hypothesis test.
- Open the Critical t tab. It is selected when the page loads.
- Choose a confidence level. The buttons cover 80%, 90%, 95%, 98%, 99% and 99.9%, and the small badge beside the label shows the matching α. For anything else, tap Custom and type either the confidence percentage or α. The two boxes stay linked, so typing one updates the other.
- Enter the degrees of freedom. Type a value (decimals are allowed, which helps with Welch degrees of freedom), nudge it with the − and + buttons, drag the slider, or press the infinity button to use the normal distribution. If you only know the sample size, switch to from n or from n1, n2.
- Pick the tails. Two-tailed splits α into two halves, one on each side. Left tail and Right tail put all of α on one side.
- Read the result. The large number is the critical t value, with a one-line summary underneath, such as “95% confidence, df = 10, two-tailed”.
A few things in this tab are easy to miss:
- The curve. The shaded red areas are the rejection regions. The dashed amber curve is the normal distribution, so you can see how far the t curve departs from it at your df.
- The three small boxes. They show the area in each tail, the equivalent normal z value, and how much wider the t value is than z, as a percentage.
- Quick table. On larger screens, a small table appears for your current df with the 90%, 95%, 99% and 99.9% values, both two-tailed and one-tailed. Tap any value and the calculator loads it.
- Drag the line. Grab the amber handle on the curve and slide it. α and the confidence level update as you move, which is a good way to build intuition.
- Show working. This opens a short step-by-step explanation, ending with a check that solves the probability back again to confirm the answer.
- Copy result, Copy APA and decimals. The first copies the result as text. The second copies a line formatted for an APA-style report. The button on the right cycles between 3, 4 and 6 decimal places.
Mode 2: p-value (the p value from t score calculator)
Use this tab when you already have a t statistic, from software, a textbook or your own working, and you want the p-value.
- Open the p-value tab.
- Type your t statistic. Negative values are fine. The − and + buttons move it in steps of 0.1.
- Set the degrees of freedom, in any of the three ways described above.
- Choose two-tailed, left tail or right tail.
- Pick your significance level α from 0.10, 0.05, 0.01 or 0.001.
- Read the p-value in the large number. The coloured bar underneath says whether the result is significant at your chosen α.
The three small boxes show the two-tailed p-value, the one-tailed p-value for the side you chose, and the critical t value at your α, so you can compare your statistic against the cut-off in the same glance. Very small p-values appear in scientific notation, for example 5.27 × 10−30, so you never see a misleading “0.0000”. You can also drag the line on the curve to try different t values.
Mode 3: t-test (one sample, Welch and pooled t test calculator)
This tab runs a complete t test from summary statistics. You do not need the raw data, only the mean, standard deviation and sample size of each group.
- Open the t-test tab.
- Choose the test type. One-sample compares one group with a target value. Welch compares two independent groups without assuming equal variances. Pooled is the classic equal-variance Student test, the one many people call the independent samples t test calculator or two sample t test calculator.
- Enter the data. Fill in the mean, standard deviation and n for the sample, and for the second group if you chose a two-sample test.
- Enter the hypothesised value. For a one-sample test this is the target mean μ0. For a two-sample test it is the hypothesised difference, which is normally 0.
- Choose the tails and α.
- Read the results. The large number is the t statistic. Underneath you will see the p-value and degrees of freedom. The three boxes show the degrees of freedom, the confidence interval and Cohen’s d, and the bar below them gives the decision.
Some details worth knowing. Welch degrees of freedom are usually not whole numbers, and the calculator keeps the decimals instead of rounding them. When you pick a left or right tail, the confidence interval becomes one-sided, with an infinity symbol on the open side. That is the same convention statistical software such as R uses. The Show working button lists the formulas with your numbers filled in, and Copy APA produces a line such as t(38.39) = 1.78, p = .083. If something is wrong, for example a standard deviation of zero or a sample size below 2, the field turns red and a message explains what to fix.
Extra features worth knowing
- Infinite degrees of freedom. The infinity button switches to the normal distribution. It is also used automatically for extremely large df, where the two are indistinguishable.
- Decimal degrees of freedom. Useful if you already have a Welch value from another program.
- Thousands separators. You can type 1,234 and the calculator reads it as 1234.
- Phone friendly. The layout shrinks to fit small screens, and tapping a tab, a tail option or Reset scrolls the calculator back into view.
- Runs in your browser. The calculations happen on your device.
How to find the t value in statistics: printed table or calculator?
The traditional method uses a t table. Take a 95% confidence interval with 14 degrees of freedom as an example. You find the row for df = 14, the column for a two-tailed α of 0.05 (which is the same as a one-tailed 0.025), and read off 2.145. With the calculator you open Critical t, choose 95%, enter 14, leave it on two-tailed, and get ±2.145. Same answer, a few seconds less, and no chance of reading the wrong row.
Where the calculator really pays off is when your numbers do not sit neatly on a printed row. Tables usually jump from 30 to 40, 60 and 120 degrees of freedom, and they only list a handful of confidence levels.
| Situation | Printed t table | This calculator |
|---|---|---|
| df = 47 | No row, so you round or interpolate | Exact value |
| Welch df = 38.39 | Not listed | Exact value |
| α = 0.037 | Not listed | Type it in Custom |
| Checking your work | Re-read the table | Show working |
| Seeing the tails | Not shown | Live curve |
If you are wondering how to calculate the t critical value without a table, the idea is this. The critical value t* is the point that leaves a probability α (or α/2 for a two-tailed test) in the upper tail of the t distribution. There is no simple formula for it, so it has to be solved numerically. The calculator does this with the regularised incomplete beta function and then refines the answer until it is accurate to many decimal places.
The same engine works the other way round. Give it a t statistic and it returns the tail probability, which is how the p-value tab works.
Two-tailed p-values for t statistics from 0.5 to 4 with 10 degrees of freedom. Red bars fall below 0.05. The p-value drops quickly once t passes about 2.
One-tailed vs two-tailed t critical value
Whether you use a two tailed t critical value calculator or a one tailed t critical value calculator setting depends on the question you are asking before you look at your data. If you only care whether a mean is higher (or only whether it is lower), a one-tailed test fits. If a difference in either direction would matter, use a two-tailed test.
The red areas are the rejection regions. The same 5% of probability is placed in two halves for a two-tailed test, or all on one side for a one-tailed test.
| Alternative hypothesis | Tail to choose | Reject H0 when |
|---|---|---|
| Mean is different (≠) | Two-tailed | |t| > 2.228 |
| Mean is greater (>) | Right tail | t > 1.812 |
| Mean is smaller (<) | Left tail | t < −1.812 |
Critical values for df = 10 and α = 0.05, taken from the calculator.
One handy rule: a one-tailed test at α uses the same critical value as a two-tailed test at 2α. That is why the 90% two-tailed column of a t table matches the one-tailed 0.05 column. Whichever you choose, decide it first. Switching to one tail after seeing the data is one of the easiest ways to fool yourself.
t table vs z table: when does the difference matter?
A z table assumes the population standard deviation is known. A t table builds in the extra uncertainty that comes from estimating it. The smaller your sample, the bigger the gap. The table below compares the two-tailed 95% critical values.
| df | Critical t | Critical z | t is larger by |
|---|---|---|---|
| 2 | 4.303 | 1.960 | 119.5% |
| 5 | 2.571 | 1.960 | 31.2% |
| 10 | 2.228 | 1.960 | 13.7% |
| 20 | 2.086 | 1.960 | 6.4% |
| 30 | 2.042 | 1.960 | 4.2% |
| 60 | 2.000 | 1.960 | 2.1% |
| 120 | 1.980 | 1.960 | 1.0% |
| 1000 | 1.962 | 1.960 | 0.1% |
Two-tailed critical t values for 90%, 95% and 99% confidence. Each curve flattens towards the matching normal value as the degrees of freedom increase.
You will often read that you can switch to z once df passes 30. At 30 degrees of freedom the t value is still about 4% bigger than z, and that matters if your result sits near the boundary. With a calculator there is no reason to round the problem away.
t distribution table: critical values of t
If you prefer to see a full t distribution table, here it is. Every value was produced by the same engine as the calculator, so the table and the tool will always agree. The columns are two-tailed confidence levels. For a one-tailed test, use the column with double the tail probability: 90% for a one-tailed α of 0.05, 95% for 0.025, and 99% for 0.005.
| df | 90% | 95% | 99% | 99.9% |
|---|---|---|---|---|
| 1 | 6.314 | 12.706 | 63.657 | 636.619 |
| 2 | 2.920 | 4.303 | 9.925 | 31.599 |
| 3 | 2.353 | 3.182 | 5.841 | 12.924 |
| 4 | 2.132 | 2.776 | 4.604 | 8.610 |
| 5 | 2.015 | 2.571 | 4.032 | 6.869 |
| 6 | 1.943 | 2.447 | 3.707 | 5.959 |
| 7 | 1.895 | 2.365 | 3.499 | 5.408 |
| 8 | 1.860 | 2.306 | 3.355 | 5.041 |
| 9 | 1.833 | 2.262 | 3.250 | 4.781 |
| 10 | 1.812 | 2.228 | 3.169 | 4.587 |
| 12 | 1.782 | 2.179 | 3.055 | 4.318 |
| 15 | 1.753 | 2.131 | 2.947 | 4.073 |
| 20 | 1.725 | 2.086 | 2.845 | 3.850 |
| 25 | 1.708 | 2.060 | 2.787 | 3.725 |
| 30 | 1.697 | 2.042 | 2.750 | 3.646 |
| 40 | 1.684 | 2.021 | 2.704 | 3.551 |
| 60 | 1.671 | 2.000 | 2.660 | 3.460 |
| 120 | 1.658 | 1.980 | 2.617 | 3.373 |
| ∞ (z) | 1.645 | 1.960 | 2.576 | 3.291 |
Two-tailed critical values of the Student’s t distribution. The last row is the normal distribution (infinite df).
Using the t value calculator for a confidence interval
The most common reason to look up a critical t value is to build a confidence interval for a mean. The formula is:
Imagine a delivery firm times 16 deliveries and finds a mean of 34.2 minutes with a standard deviation of 5.8 minutes. The standard error is 5.8 ÷ √16 = 1.45. With 15 degrees of freedom, the 95% two-tailed critical value is 2.131, so the margin of error is 2.131 × 1.45 = 3.09 minutes.
| Quantity | Value |
|---|---|
| Sample mean | 34.2 minutes |
| Standard deviation | 5.8 minutes |
| Sample size | 16 |
| Degrees of freedom | 15 |
| Critical t (95%, two-tailed) | 2.131 |
| Standard error | 1.45 |
| 95% confidence interval | 31.11 to 37.29 minutes |
Now suppose the firm promised an average of 30 minutes. The t statistic is (34.2 − 30) ÷ 1.45 = 2.897, which gives a two-tailed p-value of 0.0111. That agrees with the interval: 30 lies outside the range 31.11 to 37.29, so the data are not consistent with a 30-minute average at the 5% level.
Try it yourself: open the t-test tab, choose One-sample, enter mean 34.2, standard deviation 5.8, n = 16 and μ0 = 30. You should see t = 2.897, df = 15 and a 95% interval of 31.109 to 37.291.
Worked t-test examples
Two groups, two methods
Two seminar groups sit the same exam. Group 1 has 24 students with a mean mark of 68.4 and a standard deviation of 9.1. Group 2 has 21 students with a mean of 62.9 and a standard deviation of 11.3. Is the 5.5 mark gap real, or could it be chance? Here is what the Welch and pooled tests say.
| Result | Welch | Pooled |
|---|---|---|
| t statistic | 1.782 | 1.808 |
| Degrees of freedom | 38.39 | 43 |
| Two-tailed p-value | 0.0827 | 0.0777 |
| 95% CI of the difference | −0.75 to 11.75 | −0.64 to 11.64 |
| Cohen’s d | 0.54 | 0.54 |
Both tests agree. The p-values are around 0.08, so the difference is not significant at 0.05, and both intervals include zero. That is not the same as proving there is no difference. A gap of 5.5 marks with a medium effect size (d = 0.54) is interesting, and a larger sample would settle the question. It is a good reminder that “not significant” means “not enough evidence yet”.
A published benchmark
For an independent check, we ran an example from the NIST Engineering Statistics Handbook that compares the fuel economy of 249 American cars (mean 20.14458, standard deviation 6.41470) with 79 Japanese cars (mean 30.48101, standard deviation 6.10771) using the pooled test. The calculator and the published results match.
| Result | NIST published | This calculator |
|---|---|---|
| t statistic | −12.62059 | −12.62059 |
| Pooled standard deviation | 6.34260 | 6.34260 |
| Degrees of freedom | 326 | 326 |
| Critical t, two-tailed 5% | 1.9673 | 1.9673 |
| Critical t, one-tailed 5% | 1.6495 | 1.6495 |
A p-value check
Suppose you already have t = 2.45 with 18 degrees of freedom. In the p-value tab, the two-tailed p-value is 0.0247 and the one-tailed value is 0.0124. The critical t at α = 0.05 is ±2.101, and since 2.45 is larger, the result is significant.
One sample, Welch or pooled: which t test should you pick?
| Test | Use it when | Degrees of freedom |
|---|---|---|
| One-sample | You compare one group mean with a target value | n − 1 |
| Welch | Two independent groups, and spreads or sizes may differ | Welch–Satterthwaite, often a decimal |
| Pooled | Two independent groups with a similar spread | n1 + n2 − 2 |
If you are unsure, Welch is the safer choice. It gives almost the same answer as the pooled test when the groups are similar, and it stays reliable when they are not. Many statisticians suggest using it as the default for two independent groups.
What about a paired t test, where the same people are measured twice? Subtract the two measurements for each person, then enter the mean and standard deviation of those differences into the one-sample test with μ0 = 0. The degrees of freedom will be the number of pairs minus one.
Cohen’s d: how big is the difference?
A p-value tells you whether an effect is likely to be real. It does not tell you whether it is large enough to matter. Cohen’s d expresses the difference in standard deviations, so it can be compared across studies. The calculator gives you d next to every t-test: for one sample it is (mean − μ0) ÷ s, and for two samples it is the difference in means divided by the pooled standard deviation. It works as a simple Cohen’s d calculator when you only need the effect size.
| Cohen’s d | Rough label |
|---|---|
| 0.2 | Small |
| 0.5 | Medium |
| 0.8 | Large |
These labels come from Cohen’s widely quoted guidelines. Treat them as a starting point, not a rule. In some fields a d of 0.2 is important.
Common mistakes when using a t value or p value calculator
- Picking the tail after seeing the data. Decide on one-tailed or two-tailed before you collect or look at results.
- Using the wrong degrees of freedom. A one-sample test uses n − 1, not n. A pooled test uses n1 + n2 − 2.
- Typing the standard error instead of the standard deviation. The t-test tab expects the standard deviation of the sample. Entering the standard error makes the result far too large.
- Reading the p-value as the chance the null hypothesis is true. It is the probability of data this extreme if the null were true, which is a different thing.
- Treating “not significant” as “no effect”. A small sample can miss a real difference. Look at the confidence interval too.
- Rounding the t statistic too early. Keep several decimals until the end, then round the final answer.
- Ignoring the assumptions. The t test expects independent observations and data that are roughly normal, which matters most for small samples. Strong outliers or very skewed data can mislead it.
- Running many tests and reporting only the significant one. If you test lots of comparisons, some will look significant by chance alone.
This guide and the calculator are for learning, coursework and quick checks. For clinical, legal or other high-stakes decisions, have a qualified statistician review the analysis.
Other free tools that pair well with this one
A fair comparison starts with fair groups. If you need to split people or items into two groups at random, the list randomizer generator shuffles a list so you can assign participants without bias. Once your t test shows a difference, readers usually want to know how big it is in everyday terms, and the percentage calculator turns two means into a percentage change in seconds.
If you are a student tracking marks while you study statistics, the university grade calculator UK and the grade calculator show how each result affects your overall average.
Frequently asked questions
What is a t value calculator?
A t value calculator finds numbers from the Student’s t distribution without a printed table. This one returns critical t values from a confidence level and degrees of freedom, p-values from a t statistic, and the full results of a one-sample, Welch or pooled t test.
How do I find the critical t value for 95% confidence?
Open the Critical t tab, choose 95%, enter your degrees of freedom and pick two-tailed. For example, with 20 degrees of freedom the answer is ±2.086.
What is the difference between the t value and the p-value?
The t value (or t statistic) measures how far your sample result is from the claimed value in standard errors. The p-value converts that distance into a probability using the t distribution. A bigger absolute t gives a smaller p-value.
How do I calculate degrees of freedom?
Use n − 1 for a one-sample test and n1 + n2 − 2 for a pooled two-sample test. A Welch test uses the Welch–Satterthwaite formula, which the calculator applies for you. The from n and from n1, n2 buttons also do the sums.
Should I use a one-tailed or two-tailed test?
Use a two-tailed test unless you had a good reason, before seeing the data, to look in one direction only. Two-tailed tests are the standard choice in most research reports.
When should I use Welch’s t test instead of the pooled test?
Use Welch when the two groups may have different spreads or different sizes. It is a safe default for two independent groups because it behaves almost the same as the pooled test when variances are equal.
Why is the critical t value larger than the critical z value?
Because the standard deviation is estimated from a sample, which adds uncertainty. The t distribution has heavier tails to reflect that, so you need a larger cut-off. The gap shrinks as the degrees of freedom increase.
Can I use this as a paired t test calculator?
Yes. Work out the difference for each pair, then enter the mean and standard deviation of those differences in the One-sample test with μ0 = 0 and n equal to the number of pairs.
What does a negative t value mean?
It means your sample mean is below the value you are testing against. The sign shows direction only. For a two-tailed test, the p-value depends on the absolute size of t.
Can I calculate a t score from a mean and standard deviation?
Yes. In the t-test tab, enter the mean, standard deviation and sample size, plus the value you want to compare against. The t statistic appears as the large number, next to the p-value and confidence interval.
What does the infinity button do?
It sets the degrees of freedom to infinity, which turns the t distribution into the standard normal distribution. At 95% two-tailed, that gives ±1.960.
Is this calculator accurate?
Yes. It solves the t distribution directly rather than looking up a table, and we compared it with SciPy and mpmath on more than 300 cases. It also reproduces the published worked examples from the NIST Engineering Statistics Handbook. The Show working button lets you verify each step.
Is it fine to use a t test with a small sample?
Yes, the t test was designed for small samples, provided the data are roughly normal and the observations are independent. With very small samples, check for outliers and be cautious about conclusions.
Does the calculator save my numbers?
No. The calculations run in your browser, and the calculator does not store or send what you type.
Add this tool to your website
Want the calculator on your own site, blog or course page? Copy the code below and paste it into the HTML of any page that allows custom code, such as WordPress (Custom HTML block), Blogger (HTML view), Wix or Squarespace. The calculator appears in a frame on your page and always shows the latest version.
No comments:
Don't add any Spam link in comment box.