Define the population, event, or assumptions first
It measures straight-line association only, so a high value does not confirm a line is the right model. For x values 1 through 5 paired with their squares, the correlation is about 0.981 even though the true relationship is visibly curved. Always look at the scatterplot before trusting the number.
What each quantity represents in Pearson correlation
Name the experiment or data-generating process, distinguish sample information from population parameters, and state the event or estimate in words. Rebuild missing background in the statistics subject guide, then use statistics practice to retrieve the relationship without being told its name.
Interpret a complete Pearson correlation calculation
The numerical line is only the middle of the answer; connect it back to the event, data, or uncertainty:
Compute without losing the statistical question
- Name the unknown, and the unit the answer has to come out in.
- Match the symbols to your values. $r$ is unitless and lies between $-1$ and $1$.
- Check the conditions before substituting. Both variables must vary; correlation does not establish causation and describes linear association.
- Substitute, keep exact values to the last line, then test the sign, size, and unit against a rough estimate - The check that catches most statistics slips.
Check the range, direction, and interpretation
After using Pearson correlation, repeat the calculation with units attached and test the result against a boundary case or a quick numerical estimate.
An inference this formula does not justify
What goes wrong: Treating a strong correlation as evidence that one variable causes the other.
Why it fails: Two variables can move together because a third one drives both, or by pure coincidence in a small sample. Correlation orders the association; it says nothing about the direction of any influence.
Repair: Ask what else could explain the pattern, and remember that only a controlled experiment with randomised assignment supports a causal claim.
Questions about interpreting Pearson correlation
What does r squared add that r does not?
It converts the association into a share of variation accounted for. A correlation of 0.8 sounds like most of the story but corresponds to 0.64, meaning 64 percent of the variation in y is explained.
Does swapping which variable is x change the value?
No, the formula is symmetric in the two variables. That is a clue that correlation cannot indicate direction of cause, since it cannot even tell the two roles apart.
How strong does r have to be to matter?
It depends entirely on the field. In physics 0.9 might signal a problem with the equipment, while in social research 0.3 can be a genuinely important finding worth reporting.
Can r be exactly 1 for data that is not a straight line?
No. The value 1 is achieved only when every point sits exactly on a rising line, which is why an exact 1 in real data usually means one variable was computed from the other.
Change one input, predict whether the output should rise, fall, or stay fixed, and only then recalculate.