Lesson 9: Investigating sleep: a correlational study and Pearson's r
Interpret a correlation that is not significant
SkillCorrelation and Pearson's r
Worked example
Ten students recorded their exercise and sleep for a week.
| Exercise (min per day) | 0 | 15 | 20 | 30 | 30 | 45 | 60 | 60 | 75 | 90 |
|---|---|---|---|---|---|---|---|---|---|---|
| Sleep (h per night) | 7.2 | 7.8 | 6.9 | 7.5 | 8.1 | 7.0 | 7.9 | 7.4 | 8.3 | 7.6 |
Software gives and .
(a) Describe the correlation. (b) State whether the null hypothesis should be rejected, and conclude. (c) The students conclude that "exercise has no effect on sleep". Evaluate this conclusion.
Step 1: Direction and strength
is positive, and its size lies between 0.25 and 0.5: a weak positive correlation. Students who exercised more tended to sleep slightly longer.
Step 2: Significance
is greater than 0.05, so the correlation is not significant: retain the null hypothesis. A correlation this strong could easily arise by chance in a sample of 10.
Step 3: Evaluate "no effect"
The study cannot show an effect either way. A correlational design manipulates nothing, so it cannot test cause. And not finding a significant correlation is not proof of no relationship: with only 10 students, a real weak correlation may be missed (a possible type II error). A larger sample is needed.
Retaining H₀ means the evidence is not strong enough; it does not show that H₀ is true.
Answer
(a) A weak positive correlation (). (b) : retain H₀; there is no significant correlation between exercise and sleep in this sample. (c) The conclusion overreaches: a correlation cannot show an effect, and a non-significant result from 10 students may miss a real weak relationship (type II error).
Try a similar one
A fresh question of the same kind, checked as you answer, with hints and a full solution.
Method
Direction from the sign, strength from the size of .
Compare with 0.05: below, reject H₀; above, retain it.
Never conclude cause; a non-significant result is not proof of no relationship.
Watch out for
Saying "there is no correlation" when . There is a weak one in the sample; it is just not significant.
Concluding cause from a correlation, in either direction.