Hypothesis Test for Correlation Explained with Worked Examples

To test , compare the sample correlation with the critical value for : reject if is greater than the critical value.

Write 2x or 2*x to multiply, ^ for powers, and ; between simultaneous equations. You can also type factor, expand, derivative of, integrate … from a to b, limit … as x->0 or 15% of 80.

Try an example

Before you start

This page is easier if you are comfortable with:

What Hypothesis test for correlation means

The population correlation coefficient is written . The null hypothesis says there is no linear correlation in the population.

Use a table of critical values for the product moment correlation coefficient: it gives the minimum needed for significance at a given level and sample size.

Conclude in context, and remember that correlation does not prove causation. The coefficient is the product moment correlation coefficient, the line is the linear regression line, and the same logic appears in the binomial test and the one-sample t statistic.

This page is part of Hypothesis tests and confidence intervals in Statistics. Topics that work alongside it: Product moment correlation coefficient, Linear regression line, Hypothesis test for a binomial proportion, One-sample t statistic, Confidence interval for a mean (σ unknown) and Cumulative frequency: median and quartiles.

What is it used for?

This test uses the sample correlation coefficient to decide whether there is evidence of a real linear relationship in the population.

How to use hypothesis test for correlation

  1. State and .
  2. Calculate .
  3. Look up the critical value for and the significance level.
  4. Compare with it and conclude.

Graph

Data with product-moment correlation r = 0.904centre of the data0123456012345678Data with product-moment correlation r = 0.904xyline of best fitGraph with x from 0 to 6 and y from 0 to 8. Shows data (5); the curve line of best fit; point centre of the data at (3, 5).

With 5 points r = 0.904 is compared with the critical value for the test.

Worked examples

These examples are worked by hand, step by step, so you can compare your own working.

Example 1: a significant correlation

For pairs, . The 5% two-tailed critical value is 0.632. Test .

Worked example
Answer
Reject H0\text{Reject }H_{0}
  1. Compare
    Since , reject .
  2. Conclude
    There is evidence at the 5% level of a linear correlation.

Example 2: not significant

For pairs, . Test at the same level.

Worked example
Answer
Do not reject H0\text{Do not reject }H_{0}
  1. Compare
    Since , do not reject .
  2. Conclude
    There is not enough evidence of linear correlation.

Example 3: a negative correlation

For pairs, . The 5% two-tailed critical value is 0.632. What do you conclude?

Worked example
Answer
significant negative correlation\text{significant negative correlation}
  1. Compare the size of r with the critical value
  2. Conclude
    There is significant evidence of a (negative) linear correlation.

Special cases

A few situations come up again and again. Knowing them saves time and helps you spot answers that cannot be right.

n = 10, 5% two-tailed critical 0.6319

r = 0.6

0.6 − 0.6319 = −0.0319 (not significant).

n = 8, critical 0.7067

r = 0.75

0.75 − 0.7067 = 0.0433 (significant).

t form: r = 0.6, n = 10

t = r√(n − 2)/√(1 − r²)

2.1213.

degrees of freedom

n − 2

8.

Word problems

To test H₀: ρ = 0, compare |r| with the critical value for n (from tables). Equivalently t = r√(n − 2)/√(1 − r²) with n − 2 degrees of freedom.

n = 10

r = 0.6 versus 0.6319.

  1. 0.6 − 0.6319
  2. Check: negative, so not significant ✓-0.0319

n = 8

r = 0.75 versus 0.7067.

  1. 0.75 − 0.7067
  2. Check: positive, so significant ✓0.0433

t form Higher

r = 0.6, n = 10.

  1. 0.6 × 2.828 / 0.8
  2. Check: ✓2.1213

d.f. Higher

n − 2.

  1. 10 − 2
  2. Check: ✓8

Common mistakes

Using r not |r|

signed ✗

|r| ✓

Two-tailed.

d.f. = n − 1

n − 1 ✗

n − 2 ✓

Two parameters.

Significant means strong

strong ✗

unlikely by chance ✓

Interpretation.

Causation

causal ✗

association ✓

Interpretation.

Exam tips

  1. H₀: ρ = 0.
  2. Compare |r| with the critical value.
  3. One- or two-tailed.
  4. d.f. = n − 2 for the t form.
  5. Conclude in context.

Practice questions

Try each one on paper, then open the answer. Use the solver above to see the full working.

1. |r| − critical: r = 0.8, crit 0.7545 Foundation

0.0455

2. t: r = 0.5, n = 12, 4 dp Foundation

1.8257

3. d.f. for n = 15 Foundation

13

4. t: r = 0.8, n = 6, 4 dp Higher

2.6667

5. |r| − critical: r = 0.4, crit 0.5760 Higher

-0.176

6. t: r = 0.9, n = 11, 4 dp Higher

6.1942

Practice generator

Unlimited fresh questions with instant marking. Choose a level, work it out on paper, then type your answer.

Solve: …

Key terms

Correlation test
Tests ρ = 0.
Critical value
From tables.
Population correlation (ρ)
True value.
Significance level
α.
Degrees of freedom
n − 2.
Two-tailed
ρ ≠ 0.

Frequently asked questions

How do I test for correlation?

Compare |r| with the critical value for n.

What are the hypotheses?

H₀: ρ = 0 against H₁: ρ ≠ 0, ρ > 0 or ρ < 0.

What is the t form?

r√(n − 2)/√(1 − r²).

Does significance imply causation?

No.

Which page covers r?

product moment correlation.

Which page covers regression?

linear regression line.

More in Statistics and the formula sheets.