On this concept: Formula

Not sure which lesson to do first? Find your level and get a plan →

Math · Statistics & Probability · Grade 6-8

Correlation

Correlation describes the direction and strength of association between two numerical variables.

−1≤r≤1

Quick Answer

Correlation describes the direction and strength of association between two numerical variables. Use it when each person or thing has two measurements, such as hours of sleep and a test score, or the day's temperature and ice cream sales, and you want to say whether they rise together, move in opposite directions, or show no pattern. The correlation coefficient r runs from −1 to 1: the sign gives the direction and the size says how tightly the points follow a straight line. It never proves that one causes the other.

Why This Matters

Correlation puts a number on a pattern you can see: r=0.9 means the dots hug a rising line, r=−0.2 means a weak, slightly falling cloud. Without it, people call any upward drift 'strong'. Knowing its limits matters as much: ice cream sales and sunburns are correlated because both rise in hot weather, not because ice cream causes sunburn.

Intuitive Explanation

For each student in a music class, record the hours spent practicing in a week and the score at the recital, then plot one dot per student. If the dots drift up from left to right, more practice tends to go with higher scores: a positive correlation. If they drift down, as with outdoor temperature and a home's heating bill, the correlation is negative. If they form a shapeless cloud, as with shoe size and spelling score, there is little or no correlation.

Direction is only half the story; strength is how tightly the dots follow a line. Picture two classes whose dots both rise. In one, every dot sits close to a straight line, so a student's practice time predicts their score well, and r might be 0.9. In the other the dots rise but scatter widely, and r might be 0.4. A value of exactly 1 or −1 means every dot lies on a line; 0 means no straight-line pattern at all.

Correlation is not causation. Fires where more fire trucks show up tend to do more damage, but the trucks do not cause the damage: bigger fires bring more trucks and also do more damage. A third factor can make two things rise together. Correlation also measures only straight-line patterns: a thrown ball's height plotted against time makes a perfect arch, yet its r can be close to 0.

Think of it like this. Two correlated variables are like two dancers. With positive correlation, when one steps forward the other usually steps forward too; with negative, one steps back as the other steps forward. The tighter they stay in step, the closer r is to 1 or −1. Dancing in step does not tell you who is leading, or whether both are just following the music.

Core idea. Correlation describes how two variables move together; it does not by itself prove cause.


When to Use

Use correlation when each person or thing has two numerical measurements and the question asks whether they go together, in which direction, and how strongly. Draw the scatter plot first; correlation summarizes what it shows. If the data are categories (favorite color, type of pet), a two-way table fits instead. If the question asks whether one thing causes the other, correlation alone cannot answer it; that needs an experiment.

Tip. Ask: Do the dots show a direction and tightness of pattern?

How to Recognize It

Before describing a correlation, check that each person or thing was measured twice, with numbers, and decide what claim the question makes about the two measurements.

  1. Were both measurements taken on the same people or things, such as each student's sleep and that same student's score?

    If yes, the pairs can be correlated. If you only have one list of numbers, there is nothing to pair; describe it with a center and a spread, or a histogram, instead.

  2. Does the question claim that one variable makes the other change?

    If yes, it is a causation claim, and a correlation cannot settle it; that needs a controlled experiment. If it only asks whether the two tend to rise or fall together, correlation answers it.

  3. Do the dots bend into a curve, like a U or an arch, rather than lying in a straight band?

    If yes, r can be near 0 even though the variables are closely linked, because r only measures straight-line trends. Describe the curve in words instead. If the band is straight, r summarizes it well.

  4. Could a third quantity, like temperature or age, be pushing both variables?

    If yes, the correlation may come from that hidden variable: ice cream sales and sunburns rise together because both follow hot weather. Name the possible third variable before suggesting that either one causes the other.

Correlation compared with Causation and Scatter plot

Correlation, causation and the scatter plot all describe two linked measurements, and people slide from one to the next without noticing. Ask whether you are drawing the data (scatter plot), summarizing how the two move together (correlation), or claiming that one makes the other change (causation).

Correlation

Meaning
Use this when paired numerical data show a trend whose direction and strength need to be described.
Key test
Do the dots show a direction and tightness of pattern?
Formula
−1≤r≤1
Example
A scatter plot shows hours studied and test score. The dots rise from left to right and cluster near a line. Describe the correlation.

Causation

Meaning
One variable directly produces change in another.
Key test
Use only with evidence beyond correlation.
Example
A controlled experiment

Scatter plot

Meaning
The graph showing paired data.
Key test
Use to visualize before describing correlation.
Example
Dots on x-y axes

Formula & Notation

−1≤r≤1
r=1n−1∑i=1n(xi−xˉsx)(yi−yˉsy) where −1≤r≤1

How to read it: r summarizes direction and strength for a roughly linear association.

Worked Examples

Example 1 — Positive association

Easy

A scatter plot shows hours studied and test score. The dots rise from left to right and cluster near a line. Describe the correlation.
Show the steps and the answer
  1. Read the direction: the dots rise from left to right.

    More study hours go with higher scores, so the correlation is positive.

  2. Read the strength: the dots cluster near a line.

    Tight clustering means study time predicts score well, so r is close to 1, around 0.8 or more.

  3. Combine the two: strong and positive.

    A description of a correlation needs both its direction and its strength.

  4. Compare with a looser pattern: if the dots rose but spread in a wide band, it would be weak positive, with r nearer 0.3.

    The same direction can come with very different strengths.

Answer: Strong positive correlation

Takeaway: Describe a correlation with its direction (the sign) and its strength (how tightly the dots follow a line).

Example 2 — Cause claim

Standard

Can the scatter plot alone prove that studying caused every score increase?
Show the steps and the answer
  1. Separate what the plot shows from the claim: students who studied more tended to score higher.

    That is an association between two measured variables, nothing more.

  2. List other explanations: students who study more may also sleep more, attend more classes or already like the subject.

    A third factor could raise both study time and score.

  3. Notice that 'every' is too strong as well: even with a strong correlation, some students who studied 5 hours score lower than some who studied 2.

    Correlation describes a tendency across the group, not a rule for each person.

  4. Name what could test cause: an experiment that randomly assigns students to study 1 hour or 4 hours.

    Random assignment balances the other factors, so a difference in scores can be traced to study time.

Answer: No

Takeaway: A scatter plot can show that two variables move together; proving that one causes the other needs a controlled experiment.

Example 3 — Find the mistake: upward is not automatically strong

Application

Five students record (hours of sleep, quiz score): (6, 70), (7, 92), (8, 65), (9, 88), (10, 80). Raj writes: "The scores go up overall, so there is a strong positive correlation." Find the mistake and correct it.
Show the steps and the answer
  1. Check the direction: from 6 to 10 hours, the scores start at 70 and end at 80.

    There is a slight overall rise, so Raj is right that any trend is upward.

  2. Check the tightness: as sleep increases, scores jump to 92, drop to 65, rise to 88, then fall to 80.

    Strength depends on how closely the dots follow a line, and these zig-zag far above and below any line.

  3. Test a prediction: the student with 8 hours scored lowest of all, below the student with 6 hours.

    In a strong positive correlation, more sleep would almost always go with a higher score.

  4. Confirm with a calculator: r≈0.22.

    An r near 0.2 is a weak positive correlation, far from the 0.8 or more that 'strong' suggests.

Answer: Weak positive correlation (r≈0.22), not strong.

Takeaway: An upward trend is not automatically strong; strength is how tightly the dots follow a line.

Where You'll See It

  • Ice cream sales on hot days

    A beach shop finds that daily ice cream sales rise with the temperature, a strong positive correlation, and uses the forecast to plan its stock. Sunburns also rise on hot days, so ice cream sales and sunburns are correlated too, but the heat drives both.

  • A used car's age and price

    Used cars show a negative correlation between age and price: each extra year tends to lower the value. Car-pricing websites use that pattern to estimate a fair price, then adjust for mileage and condition, which explain much of the scatter around the trend.

  • Height and arm span

    Measure classmates' heights and arm spans and the dots fall almost on a line, with r close to 1, because most people's arm span is about equal to their height. That tight correlation lets you estimate one measurement from the other.

Common Mistakes

  1. The slip: Saying correlation proves causation

    Instead: association alone is not proof of cause.

  2. The slip: Calling any upward pattern strong

    Instead: strength depends on how tightly dots cluster around the trend.

  3. The slip: Using correlation with categorical data

    Instead: correlation needs paired numerical variables.


Mini Practice

Try these on your own. Tap Reveal when you want to check.

  1. Match each value to a description: r=−0.9, r=0.1, r=0.75. Descriptions: almost no linear relationship, strong negative, fairly strong positive.

    Hint: The sign gives direction; closeness to 1 or −1 gives strength.

  2. Temperatures of 15,20,25,30,35 degrees Celsius go with ice-cream sales of 120,150,210,260,300 cones. Is the correlation positive, negative or close to zero?

    Hint: As one list goes up, what does the other list do?

  3. Hours of TV per day 1,2,3,4,5 go with test scores 90,85,78,70,66. Describe the direction of the correlation.

    Hint: Compare the score at 1 hour with the score at 5 hours.

  4. Which shows a stronger linear relationship: r=−0.82 or r=0.64?

    Hint: Ignore the sign when judging strength.

  5. In one town, months with more ice-cream sales also have more drownings (r=0.8). Does this show that ice cream causes drowning?

    Hint: Is there a third thing that could make both numbers go up?

  6. Dots on a scatter plot drift upward but are widely scattered, and r=0.3. Leo says: "The dots go up, so it is a strong positive correlation." Correct him.

    Hint: Strength is about how tightly the dots hug a line.

Want the full set?

50 practice questions for this concept — free to try, every one with a complete worked solution showing the why, not just the answer.

Frequently Asked Questions

Is a correlation of −0.8 weaker than one of 0.5?

No, it is stronger. The sign only tells the direction: negative means one variable tends to fall as the other rises. The size, ignoring the sign, tells how tightly the points hug a line, and 0.8 is closer to 1 than 0.5 is. So a tight downhill cloud with r=−0.8r = -0.8 shows a stronger relationship than a loose uphill one with r=0.5r = 0.5.

Why do some silly correlations look so strong?

Because when you compare enough pairs of variables, some will line up by chance. In one famous example, for about a decade in the US, the number of films Nicolas Cage appeared in each year rose and fell with the number of people who drowned by falling into swimming pools. Nobody thinks one causes the other. With only ten or so data points and thousands of pairs to check, a few strong but meaningless correlations are bound to appear.

Does the correlation change if I swap the two axes?

No. Correlation treats the two variables the same way, so plotting score against sleep or sleep against score gives the same rr. What does change when you swap the axes is the line of best fit used for predictions. That is why making a prediction means choosing which variable you are predicting from and which one you are predicting.

Learning Path

Correlation

You are here

You need the Mean and Standard Deviation, since r measures how two variables sit above and below their means together. Correlation adds one number for the direction and strength of a straight-line pattern in paired data. The Scatter Plot shows the pattern that r summarizes, and Inference for Regression asks whether a pattern seen in a sample holds for the whole population.

See Also