What is a scatter diagram?
A scatter diagram plots paired observations, for example the cure temperature and the adhesion strength of each batch, as points on two axes. It is one of the seven basic quality tools and the quickest way to test a suspected cause-and-effect relationship from a fishbone diagram with real data.
Reading a scatter diagram
- Upward cloud: positive correlation. Y rises as X rises.
- Downward cloud: negative correlation. Y falls as X rises.
- Round, shapeless cloud: no linear relationship.
- Curved pattern: a non-linear relationship. r understates it.
- Separate clusters: different groups (machines, materials); use the group column or stratification.
Scatter diagram example
The example plots 40 batches of cure temperature against adhesion strength from two coating lines. Strength rises with temperature, a strong positive correlation (r ≈ 0.83 overall). But line B sits about 2 N lower than line A at the same temperature, and within each line the correlation is even stronger (r ≈ 0.92), because mixing the two lines adds scatter. That points to a second factor specific to line B, exactly the kind of insight the group column reveals.
Common pitfalls
- Correlation is not causation. Confirm with an experiment.
- Narrow ranges hide relationships. If X barely varies in your data, the effect cannot show.
- Outliers distort r. Investigate them before deleting them.
- Extrapolating the trend line beyond the data range is risky.
How to use this Scatter Diagram tool
- Choose the two variables. Put the suspected cause on X and the effect on Y, and name both with their units.
- Enter or paste the pairs. Paste two columns (X and Y) from Excel, plus an optional group column such as machine or line.
- Look at the pattern. Check the direction and tightness of the point cloud, and look for outliers or curved patterns.
- Read r and R². The tool calculates the correlation, R², the trend line and the slope.
- Check the groups. Compare correlations by group to see whether stratification changes the picture.
- Confirm and export. Confirm the relationship with an experiment or process knowledge, and export the chart and statistics.
Frequently asked questions
What is a scatter diagram?
A scatter diagram (scatter plot) shows pairs of numerical data as points, with one variable on each axis. It reveals whether the two variables are related, how strongly, and in which direction.
What does the correlation coefficient r mean?
The Pearson correlation r ranges from -1 to +1. Values near +1 mean a strong positive linear relationship, near -1 a strong negative one, and near 0 no linear relationship. As a rough guide, |r| of 0.7 or more is strong, 0.4 to 0.7 moderate and 0.2 to 0.4 weak.
What is R²?
R² (the coefficient of determination) is the square of r. It tells you what share of the variation in Y is explained by a straight-line relationship with X. An R² of 0.64 means 64% of the variation in Y is associated with X.
Does correlation prove causation?
No. Two variables can move together because a third factor drives both, or by coincidence. Use the scatter diagram to find candidates, then confirm cause and effect with process knowledge or a designed experiment.
Which variable goes on which axis?
Put the suspected cause (the input you can control, such as temperature or pressure) on the X axis and the effect (the output, such as strength or defect rate) on the Y axis.
Why add a group column?
Groups such as machines, lines or suppliers are shown in different colours. A relationship that looks weak overall can be strong within each group, or two groups may sit on different levels, which points to a stratification factor.