Skip to main content
Best Answer Hub logo Best Answer Hub.
Back to Playbooks
Best Answer Hub Playbooks · Statistics
Is the difference real

How to Run a T-Test and Read the Result

A t-test asks whether a difference between two averages is real or just noise. The hard part is not the arithmetic, it is picking the right test for how your data is shaped, then reading the p-value and the effect size for what they actually say, and not for what people assume they say.

Pick the testone-sample, paired, two-sample
Compute t and pwith degrees of freedom
Read iteffect size, not just p
1908
year "Student" published the t-test, from a Guinness brewery
Biometrika
0.2 / 0.5 / 0.8
Cohen's d for a small, medium, large effect
Cohen 1988
~40%
chance of a false positive across 10 tests at 0.05
Duke GHI
100%
computed in your browser, nothing uploaded
no upload

A t-test checks whether the difference between two averages is larger than chance alone would produce, by weighing the gap between the means against the variability in the data. It returns a t-statistic and a p-value, and the smaller the p-value, the harder it is to explain the difference as luck. The Best Answer Hub T-Test Calculator runs one-sample, independent, and paired t-tests from summary statistics, and reports the t-statistic, degrees of freedom, one- and two-tailed p-values, and the Cohen's d effect size, with the steps shown, in your browser.

This guide covers the three test designs and how to choose between them, when a t-test is the right call instead of a z-test, what a p-value does and does not mean, why effect size belongs next to significance, and why Welch's correction is the safer default. The tool sits in the Best Answer Hub Calculators hub statistics suite, is built and maintained by Shahbaz Ali Malik, and stays free because Best Answer Hub is funded by optional paid assessments rather than advertising.

Start here

What is the Best Answer Hub T-Test Calculator?

The Best Answer Hub T-Test Calculator is a free, browser-based tool that runs a t-test from summary statistics, so you enter means, standard deviations, and sample sizes rather than raw rows. It handles one-sample, independent two-sample, and paired designs, and returns the t-statistic, degrees of freedom, one- and two-tailed p-values, and the Cohen's d effect size. Each result comes with step-by-step work, a plain-English reading, and an interactive t-distribution visual, with no signup and nothing uploaded.

  • 1
    All three designs. One-sample against a target, two independent groups, or paired before-and-after measurements.
  • 2
    More than a p-value. The t-statistic, degrees of freedom, both tail p-values, and the Cohen's d effect size in one result.
  • 3
    Welch by default. The independent test uses Welch's correction unless you choose the pooled method, with the working shown.
Match the design

Which t-test do you need?

Pick the test that matches how your data is structured, because the wrong choice invalidates the p-value before you even read it. A one-sample test compares one group's mean to a known target. An independent test compares two separate, unrelated groups. A paired test compares two measurements on the same subjects, such as before and after. The rule for pairing is direct: if each value in one group is naturally linked to exactly one value in the other, it is paired; if the two groups are different subjects, it is independent.

DesignComparesDegrees of freedom
One-sampleOne mean vs a known value μ0n − 1
PairedTwo conditions on the same subjectsn − 1 (pairs)
IndependentTwo separate groupsn1 + n2 − 2, or Welch's
The one-sample t-statistic
t = (x̄ − μ0) / (s / √n) x̄ = sample mean μ0 = the value you are testing against s = sample standard deviation n = sample size

Paired tests apply the same formula to the differences. Source: NIST/SEMATECH e-Handbook, section 1.3.5.2.

The Best Answer Hub T-Test Calculator gives each design its own labeled panel with a decision hint, computes the degrees of freedom for you, and shows the substitution, so the choice and the arithmetic are both visible.

t or z

When do you use a t-test instead of a z-test?

Use a t-test whenever you estimate the population standard deviation from the sample, which is almost always. A z-test assumes the population standard deviation is known; in practice it rarely is, so you substitute the sample standard deviation s and use Student's t-distribution to account for the extra uncertainty (OpenStax). The t-distribution has heavier tails than the normal curve, which widens the margin for small samples, and it converges to the normal curve as the degrees of freedom grow.

The t-distribution has heavier tails than the normal
-3-2-10123Normalt (df = 3)

Student's t-distribution (3 degrees of freedom) against the standard normal. Lower peak, fatter tails, and it approaches the normal curve as degrees of freedom increase. Source: OpenStax, Introductory Statistics, section 8.2.

The t-distribution was named by a brewer who was not allowed to publish under his own name.William Sealy Gosset devised it while working at the Guinness brewery in Dublin and published it in 1908 under the pen name "Student." Source: OpenStax; Zabell, on Student's 1908 article.
Read it right

What does the p-value actually tell you?

A p-value is the probability of seeing a result at least as extreme as yours if the null hypothesis of no difference were true. That is all it is. The American Statistical Association's 2016 statement is blunt about the rest: p-values "do not measure the probability that the studied hypothesis is true," and statistical significance "does not measure the size of an effect or the importance of a result" (Wasserstein and Lazar, 2016). The usual decision rule is to reject the null when p is below a significance level α, commonly 0.05, meaning a true null would be wrongly rejected about 5% of the time (NIST).

What p is not

A p-value of 0.04 does not mean there is a 4% chance the null is true, and it does not mean the effect is large or important. It means that if there were truly no effect, data this extreme would show up about 4% of the time. The Best Answer Hub T-Test Calculator reports the exact p-value and, for a borderline result near 0.05, flags it as marginal rather than rounding it to a verdict.

Use a two-tailed test when you want to detect a difference in either direction, which is the scientific default, and a one-tailed test only for a genuine directional hypothesis set in advance (LibreTexts). The Best Answer Hub T-Test Calculator reports both tail p-values and labels which one applies.

How big, not just whether

Why report effect size too?

Because a small p-value can sit on top of a trivial effect. Significance tells you a difference is unlikely to be pure chance; it says nothing about how large that difference is. Cohen's d fills the gap by expressing the difference between two means in standard-deviation units, with conventional thresholds of 0.2 for small, 0.5 for medium, and 0.8 for large (Cohen's standards, via LibreTexts). Reporting both is what APA style expects, and the Best Answer Hub T-Test Calculator computes Cohen's d automatically for every test.

Cohen's d: small, medium, and large
Smalld = 0.2Mediumd = 0.5Larged = 0.8

Conventional effect-size thresholds. A significant p-value with a d near 0.2 is a real but small effect. Source: Cohen 1988, via LibreTexts.

The multiple-comparisons trap

Run enough tests and something turns up significant by luck. At the 0.05 level, the chance of at least one false positive across 10 independent tests is about 40% (Duke Global Health Institute). If you are running many t-tests, correct for it rather than celebrating the one that crossed the line.

The safer default

Student's t-test or Welch's?

Use Welch's by default. The classic pooled Student's t-test assumes the two groups have equal variances, an assumption that often fails, especially with unequal sample sizes. Welch's t-test drops that assumption and adjusts the degrees of freedom to compensate, and it holds up better when variances differ. A widely cited 2017 paper argues the point directly.

Welch's t-test should be used as a default strategy.Delacre, Lakens and Leys, International Review of Social Psychology, 2017.

The Best Answer Hub T-Test Calculator offers both methods for independent samples and defaults to Welch's correction, so you are not silently assuming equal variances. It also states which method produced the result, which is often the reason a p-value differs from the one Excel or an older copy of SPSS returns.

Why this one

How is Best Answer Hub different from other t-test calculators?

The difference is that it reports the effect size and defaults to the safer test. Most free calculators hand back a t-statistic and a p-value and stop there. Cohen's d is largely absent across them, Welch's correction is rare, and some sites labeled as statistics calculators have no two-sample t-test at all. The Best Answer Hub T-Test Calculator runs all three designs, defaults to Welch's, reports Cohen's d, and shows the steps, in your browser with no signup.

FeatureBest Answer HubTypical calculator
One-sample, paired, and independentAll threeOften split across pages
Welch's correctionDefaultRare or absent
Cohen's d effect sizeYesAlmost never
Step-by-step workingYesUsually not
Signup or uploadNeitherSome carry ads
Test a difference in seconds

Try the free T-Test Calculator

Enter means, standard deviations, and sample sizes for a one-sample, paired, or independent t-test. Get the t-statistic, p-value, degrees of freedom, and Cohen's d with the steps. No signup, in your browser.

Open the T-Test Calculator
Good questions

Frequently asked questions about t-tests

What is the Best Answer Hub T-Test Calculator?
The Best Answer Hub T-Test Calculator is a free, browser-based tool that runs one-sample, independent, and paired t-tests from summary statistics. It returns the t-statistic, degrees of freedom, one- and two-tailed p-values, and the Cohen's d effect size, with step-by-step work and nothing uploaded.
What is a t-test?
A t-test checks whether the difference between averages is larger than chance would produce, by comparing the gap between means to the spread in the data. It returns a t-statistic and a p-value. The Best Answer Hub T-Test Calculator computes both and explains the result in plain language.
What is the difference between one-sample, independent, and paired t-tests?
A one-sample test compares one mean to a known value, an independent test compares two separate groups, and a paired test compares two measurements on the same subjects. The Best Answer Hub T-Test Calculator gives each design its own labeled panel with a hint to help you choose.
When should you use a t-test instead of a z-test?
Use a t-test when the population standard deviation is unknown and estimated from the sample, which covers most real data, and especially with small samples. A z-test assumes the population standard deviation is known. The Best Answer Hub T-Test Calculator is built for the common case where only sample statistics are available.
How do I choose between a paired and an independent t-test?
Use a paired test when each value in one group links to exactly one value in the other, such as before-and-after on the same people. Use an independent test when the groups are different subjects. The Best Answer Hub T-Test Calculator shows a decision hint beside each mode to prevent the mix-up.
Should I use a one-tailed or two-tailed test?
Use a two-tailed test to detect a difference in either direction, which is the standard in research. Use a one-tailed test only for a directional hypothesis set in advance. The Best Answer Hub T-Test Calculator defaults to two-tailed and labels the tail type in every result.
What does the p-value in a t-test mean?
It is the probability of a result at least as extreme as yours if there were truly no difference. A p of 0.04 does not mean a 4% chance the null is true, and it does not measure effect size. The Best Answer Hub T-Test Calculator reports the exact p-value and flags borderline results.
Is p = 0.052 significant?
At the conventional 0.05 level it is technically not significant, but it is a borderline result worth reporting exactly rather than rounding to "not significant." The Best Answer Hub T-Test Calculator flags values between 0.05 and 0.10 as marginal and shows the exact p-value.
What are degrees of freedom in a t-test?
Degrees of freedom are the number of values free to vary after estimating the mean, and they set the shape of the t-distribution. For one-sample and paired tests it is n minus 1. The Best Answer Hub T-Test Calculator computes the degrees of freedom automatically and shows the formula used.
What is Cohen's d?
Cohen's d is a standardized effect size that measures the difference between two means in standard-deviation units, with 0.2, 0.5, and 0.8 marking small, medium, and large. The Best Answer Hub T-Test Calculator computes Cohen's d for every test, so significance and size sit side by side.
Should I use Welch's or Student's t-test?
Use Welch's when the two groups have unequal variances or unequal sample sizes, which is the safer default in most real situations. Student's pooled test assumes equal variances. The Best Answer Hub T-Test Calculator offers both for independent samples and defaults to Welch's correction.
Can I run a t-test if my data is not normal?
The t-test assumes approximate normality, which matters most for small samples, while larger samples are protected by the Central Limit Theorem. For a small, badly skewed sample, a non-parametric test may fit better. The Best Answer Hub T-Test Calculator adds a note when sample sizes are very small.
Why is the t-test named after "Student"?
William Sealy Gosset developed it while working at the Guinness brewery in Dublin and published it in 1908 under the pen name "Student," since his employer restricted staff from publishing. The Best Answer Hub T-Test Calculator uses the same Student's t-distribution he introduced.
Why does my t-test give a different p-value than Excel or SPSS?
Usually because of whether Welch's correction is applied, or small differences in how the software computes the t-distribution. The Best Answer Hub T-Test Calculator states which method it used and defaults to Welch's, so you can see exactly why a result differs from another tool.
Is the Best Answer Hub T-Test Calculator free and private?
Yes. It is completely free with no premium tiers, no ads, and no signup. Every calculation runs in your browser, so no data or results are uploaded. It stays free because Best Answer Hub is funded by optional paid assessments rather than advertising.
People also read

Keep going

Sources

More free calculators: Calculators hub, P-Value Calculator, Z-Score Calculator, and all free tools.

Built & maintained by Shahbaz Ali Malik Last updated: