What is a Pearson r Critical Value Calculator?
An r critical value is the smallest absolute Pearson correlation that is statistically significant for n pairs at level α. This r critical value calculator finds the critical t with n − 2 degrees of freedom and converts it with r* = t* / √(t*² + n − 2). For n = 25 and a two-tailed α = 0.05, any |r| ≥ 0.3961 is significant; with fewer pairs the cut-off is higher.
What the Pearson r Critical Value Calculator does
Enter the sample size (or the degrees of freedom), α and the tail. The calculator returns the critical value of Pearson’s r, the matching critical t, the smallest significant r², and a table of critical r at six common α levels for your n. It replaces the printed r critical value table and works for any sample size, so you never have to interpolate between rows.
Use it to judge a correlation at a glance: if |r| is at least the critical value, the correlation is significantly different from zero. It is also handy before collecting data, to see how strong a correlation must be to show up with the sample size you can afford. This is a critical value tool, not an r calculator: to estimate the value of r from your data, use the Correlation Coefficient Calculator, and for the exact p-value of an observed r, use the Pearson r to P-Value Calculator.
How to use the Pearson r Critical Value Calculator
- Choose whether to enter the Sample size n (number of pairs) or the Degrees of freedom (n − 2).
- Enter the Significance level (α), usually 0.05.
- Pick the Tail: Two-tailed for any correlation, Right for a predicted positive correlation, Left for a predicted negative one.
- Optionally enter Your sample r to get a significant / not significant decision for your own correlation.
- Read the critical r and the decision rule. The step-by-step solution shows how the critical t is converted to r.
How to find the critical value of r
To find the critical value of r by hand you need three things: the number of pairs n, the significance level α and the tail of the test. Then either read a critical r value table or convert a critical t.
- Count the pairs. n is the number of (x, y) observations, not the number of variables.
- Work out the degrees of freedom. df = n − 2. Printed tables of critical values of r are usually indexed by df, not n, which is the most common lookup mistake.
- Choose α and the tail. Use two-tailed unless you predicted the sign of the correlation before seeing the data.
- Find the critical t. t* = T⁻¹(1 − α/2; df) for a two-tailed test, or T⁻¹(1 − α; df) for a one-tailed test. The T Critical Value Calculator gives it for any df.
- Convert to r. r* = t* / √(t*² + df). For a left-tailed test the critical value is −r*.
Comparing your r with the critical value
Compare the absolute value of your sample correlation with the critical r for the same n, α and tail, or type it into the optional Your sample r field above and the calculator makes the comparison for you:
- If |r| is greater than or equal to the critical value, the correlation is statistically significant: reject H₀: ρ = 0 at that α.
- If |r| is less than the critical value, do not reject H₀. The data are consistent with no linear correlation, which is not proof that ρ = 0.
- The decision always agrees with the p-value: |r| ≥ r* exactly when p ≤ α, and with the t-test, since |t| ≥ t* at the same point.
- Example: 12 students give r = 0.62 between hours studied and exam score. With df = 10, the two-tailed critical r at α = 0.05 is 0.5760. Because 0.62 exceeds 0.5760, the correlation is significant (t = 2.499, p ≈ 0.032). At α = 0.01 the critical r rises to 0.7079, so it would not be significant there.
r critical value formula
The significance test for Pearson’s r is a t-test, so the critical r is the critical t re-expressed on the correlation scale.
| Symbol | Meaning |
|---|---|
| r* | critical correlation coefficient |
| t* | critical t value with n − 2 df |
| n | number of (x, y) pairs |
| ρ | population correlation |
Derivation: square t = r√df / √(1 − r²) to get t²(1 − r²) = r²·df, collect the r² terms to get t² = r²(df + t²), and take the square root. The map from r to t is increasing, so a cut-off on t is a cut-off on r.
A quick approximation for a two-tailed test at α = 0.05 is r* ≈ 2 / √n; it is within about 0.01 of the exact value for n ≥ 20.
Sources: the t-test for a correlation is described under testing using Student’s t-distribution, and the critical t values match the NIST/SEMATECH e-Handbook t table.
Worked example: critical r for 25 pairs
Two-tailed test at α = 0.05 with n = 25
Problem. A researcher measures n = 25 pairs. How large must |r| be to be significant in a two-tailed test at α = 0.05?
- 01Degrees of freedom: df = 25 − 2 = 23.
- 02Critical t: t* = T⁻¹(0.975; 23) = 2.0687.
- 03Convert: r* = 2.0687 / √(2.0687² + 23) = 2.0687 / √27.2793 = 2.0687 / 5.2230 = 0.3961.
- 04The quick rule 2/√25 = 0.40 is close.
Answer. Any correlation with |r| ≥ 0.3961 is significant at the 5% level with 25 pairs; that is, x must explain at least r² = 0.1569 (about 16%) of the variation in y. A one-tailed test lowers the bar to r ≥ 0.3365.
Critical values of r table (two-tailed)
The cut-off falls quickly as n grows: with 5 pairs you need |r| ≥ 0.878 at α = 0.05, but with 100 pairs |r| ≥ 0.197 is enough. For a one-tailed test at α, read the two-tailed column for 2α (for example, the α = 0.10 column gives one-tailed 0.05 values). For every df from 1 to 1,000, see the full r critical value table.
| n (pairs) | df = n − 2 | α = 0.10 | α = 0.05 | α = 0.02 | α = 0.01 |
|---|---|---|---|---|---|
| 5 | 3 | 0.8054 | 0.8783 | 0.9343 | 0.9587 |
| 10 | 8 | 0.5494 | 0.6319 | 0.7155 | 0.7646 |
| 12 | 10 | 0.4973 | 0.5760 | 0.6581 | 0.7079 |
| 15 | 13 | 0.4409 | 0.5140 | 0.5923 | 0.6411 |
| 20 | 18 | 0.3783 | 0.4438 | 0.5155 | 0.5614 |
| 25 | 23 | 0.3365 | 0.3961 | 0.4622 | 0.5052 |
| 30 | 28 | 0.3061 | 0.3610 | 0.4226 | 0.4629 |
| 40 | 38 | 0.2638 | 0.3120 | 0.3665 | 0.4026 |
| 50 | 48 | 0.2353 | 0.2787 | 0.3281 | 0.3610 |
| 100 | 98 | 0.1654 | 0.1966 | 0.2324 | 0.2565 |
| 200 | 198 | 0.1166 | 0.1388 | 0.1644 | 0.1818 |
Critical r in Excel, R, Python and on a TI-84
No package has a built-in critical-r function, so each one finds the critical t and converts it. The examples use df = 23 (n = 25) and a two-tailed α = 0.05; each returns 0.3961.
| Tool | Two-tailed critical r |
|---|---|
| Excel / Google Sheets | =T.INV.2T(0.05,23)/SQRT(T.INV.2T(0.05,23)^2+23) |
| R | t <- qt(0.975, 23); t / sqrt(t^2 + 23) |
| Python (SciPy) | t = stats.t.isf(0.025, 23); t / (t**2 + 23) ** 0.5 |
| TI-84 | invT(0.975, 23) gives t*, then compute t* / √(t*² + 23) |
Common use cases
- Checking whether a correlation from a correlation calculator or spreadsheet is significant without computing a p-value.
- Planning a study: seeing the weakest correlation detectable as significant with a given number of participants.
- Screening a correlation matrix quickly, since every entry with the same n shares the same cut-off.
- Reading a correlation reported in a paper: with the sample size given, you can check whether the stated r clears the cut-off.
- Teaching the link between correlation, the t distribution and sample size.
Assumptions and limitations
- The critical value assumes independent pairs from a roughly bivariate normal population and tests only H₀: ρ = 0.
- Exceeding the critical r shows the correlation is unlikely to be zero, not that it is strong or causal. Report r and r² as well.
- The cut-off is for Pearson’s r. Spearman’s rank correlation has its own exact critical values for small n, although the t approximation is often used for n above about 10; see the Spearman Correlation Calculator.
- Screening many correlations at α = 0.05 will produce false positives; adjust α (for example Bonferroni) when testing many pairs.
Frequently asked questions
What is an R critical value?
What is the critical value of r for n = 30?
How is the critical value of r calculated?
Why does the critical r get smaller as n increases?
Is the critical r the same for positive and negative correlations?
What degrees of freedom does a correlation test use?
Is 0.05 a critical value?
What is the R value calculator?
Should I use a one-tailed or two-tailed critical r?
Cite this page
Statculator. (2026, November 14). Pearson r Critical Value Calculator. Retrieved November 14, 2026, from https://statculator.com/p-values-critical-values/r-critical-value-calculator/