Hypothesis test for correlation

Hypothesis test for correlation

Spec 5.1 · Edexcel A-Level Maths: Revision notes

Studying this course? Keep your progress in one place.

Key facts

  • The product moment correlation coefficient rr satisfies −1≤r≤1-1\le r\le1 and measures linear association only.
  • rr comes from a sample; the population correlation coefficient is ρ\rho.
  • Hypotheses use ρ\rho: H0:ρ=0H_0:\rho=0 against H1:ρ≠0H_1:\rho\ne0, ρ>0\rho>0 or ρ<0\rho<0.
  • Reject H0H_0 if rr lies in the critical region given by the critical value.
  • Conclude in context. Correlation does not imply causation.

Correlation coefficient

rr measures how closely points lie to a straight line, from −1-1 to 11.

The product moment correlation coefficient rr satisfies −1≤r≤1-1\le r\le1.

  • r=1r=1: all points on a line with positive gradient.
  • r=−1r=-1: on a line with negative gradient.
  • rr near 0: no linear association.

Find rr on your calculator. A perfect curve can still give rr near 0.

−1−0.8−0.6−0.4−0.200.20.40.60.81-1 perfect negative0 no linear relationship1 perfect positive

Which value cannot be a product moment correlation coefficient?

Sample and population

rr describes the sample; the test uses it to make a claim about the population value ρ\rho.

rr is calculated from a sample. The population correlation coefficient is ρ\rho (rho).

Hypotheses are always written in terms of ρ\rho, never rr.

  1. 1

    Sample

    calculate the product moment correlation coefficient r

  2. 2

    Hypothesis test

    compare r with the critical value for the sample size n

  3. 3

    Population

    decide whether the population correlation ρ differs from 0

From a sample correlation to a claim about the population

Sample

  • Data collected
  • Gives rr

Population

  • Whole group of interest
  • Has ρ\rho

Which is a correctly written null hypothesis for no correlation?

Stating the hypotheses

H0:ρ=0H_0:\rho=0; the alternative is one- or two-tailed depending on the question.

The null hypothesis is H0:ρ=0H_0:\rho=0. The alternative hypothesis is H1:ρ≠0H_1:\rho\ne0 (two-tailed), ρ>0\rho>0 or ρ<0\rho<0 (one-tailed).

Use one-tailed only if the direction is stated before looking at the data.

−1−0.8−0.6−0.4−0.200.20.40.60.81ρ = 0 (H₀)

Two-tailed

  • H1:ρ≠0H_1:\rho\ne0
  • 'There is correlation'

One-tailed

  • H1:ρ>0H_1:\rho>0 or ρ<0\rho<0
  • Direction expected in advance

A teacher expects more revision to give lower anxiety scores. Which H1H_1 is correct?

Using critical values

Compare rr with the critical value for your sample size and level; reject H0H_0 if it is in the critical region.

Use the table for nn and the level. One-tailed: reject if r>r> critical value (ρ>0\rho>0) or r<−r<- critical value (ρ<0\rho<0). Two-tailed: reject if ∣r∣>|r|> critical value.

If rr is in the critical region, reject H0H_0; otherwise do not.

−1−0.8−0.6−0.4−0.200.20.40.60.81-0.44380.4438r = 0.52

Worked example

n=20n=20, two-tailed, 5% level, critical value 0.44380.4438. The sample has r=0.52r=0.52.

Two-tailed test, critical value 0.44380.4438, r=−0.5r=-0.5. What do you do?

Conclusion and causation

State the decision about H0H_0 and what it means in context, never claiming proof or causation.

Alternatively use a p-value: reject H0H_0 if it is less than the significance level, such as 0.018<0.050.018<0.05.

Write whether H0H_0 is rejected and what that means in context, e.g. 'significant evidence of positive correlation between fertiliser and yield'.

Correlation does not imply causation: another variable may affect both, or the cause may run the other way.

−1−0.8−0.6−0.4−0.200.20.40.60.81−0.63190.6319

Can conclude

  • Evidence of correlation at the stated level

Cannot conclude

  • That one variable causes the other
  • That the test 'proves' anything

Fertiliser use and yield show significant positive correlation. What can you conclude?

Try an exam question

A sample of 15 pairs gives r=−0.62r=-0.62. The critical value for a one-tailed test at 5% with n=15n=15 is 0.44090.4409. Test whether there is negative correlation, and comment on causation.

[4 marks]

That's the notes covered.

Carry on to the next subtopic.

Try a question

Which pair of hypotheses should the student use?

1 mark, marked against the mark scheme

Revision notes