What does correlation actually tell you?
You see this whenever two things change together. If revision time goes up and test scores also tend to go up, there is a relationship worth noticing. Correlation is about spotting a pattern between two variables, not proving that one directly causes the other.
<b>Correlation</b> describes how strongly two sets of data are connected and in which direction they move.
A quick surprise: two variables can match really closely and still not be a cause-and-effect pair. For example, ice cream sales and sunburn cases can both rise at the same time, but buying ice cream does not cause sunburn. The weather is affecting both.
How do you describe the direction?
If your points tend to rise from left to right, you have positive correlation. If they tend to fall from left to right, you have negative correlation. If the points look scattered with no clear trend, that is no correlation.
<b>Positive correlation</b> means larger values of one variable are usually linked with larger values of the other.
<b>Negative correlation</b> means larger values of one variable are usually linked with smaller values of the other.
<b>No correlation</b> means the data does not show a clear pattern connecting the variables.
Think of it like a music app. If longer playlists usually have more total listening time, that is positive. If your phone battery percentage drops as screen time increases, that is negative. If shoe size and maths score are paired up, you would not expect a useful trend.
How do you spot it on a graph?
You usually use a scatter graph. Each point represents one pair of values. The overall shape matters more than any single point.
- Put one variable on the horizontal axis and the other on the vertical axis.
- Plot each data pair as one point.
- Look at the overall direction of the cloud of points.
- Decide whether the pattern is positive, negative, or has no clear link.
<b>Worked example</b>
A teacher records hours of revision and test score for several students. On the scatter graph, the points mostly go upward from left to right.
- The graph uses two variables: revision hours and test score.
- The points rise as you move right.
- That means students who revised more often scored higher.
- So the relationship is <b>positive correlation</b>.
This does <b>not</b> prove that revision is the only reason for the higher scores, but it does show a clear connection in the data.
When you answer questions, describe both the direction and the strength if you can. Saying only "there is correlation" is too vague.
6 more sections in this topic
Pick this up in your Library: it holds the whole topic, notes, cheatsheet and questions.