Scatter Diagram: Examples

The Quality Toolbook > Scatter Diagram > Examples

When to use it | How to understand it | Example | How to use it | Practical variations

<-- Previous | Next -->

Example

A town planning team, during an investigation of road accidents, identified a number of possible causes. Three main causes were suspected: the speed of the vehicles, the traffic density and the local weather conditions. As there was no clear evidence available to support any of these hypotheses, they decided to measure them, and used Scatter Diagrams to check whether the link between any of the causes was strong enough to take further action.

In order to get sufficient measures, they made daily measures for two months, using local road sensors and reports from the ambulance service. Scatter Diagrams were drawn for each possible cause against the accident count. The results enabled the following conclusion to be made:

• There was a low, positive correlation with traffic density.
• There was an inconclusive correlation with road conditions.
• There was a high, positive correlation with traffic speed, with accidents dropping off more sharply under 30 mph.

As a consequence, more traffic speed control measures were installed, including signs and surfaces. This resulted in a measurable decrease in accidents.

Fig. 1. Example Scatter Diagrams

Other examples

• A baker suspects that the standing time of the dough is affecting the way it rises. A Scatter Diagram of rise time against measured bread density shows a fair correlation on an inverse U-shaped distribution. He thus uses the time at the highest point on the curve to get the best chance of well-risen bread.
• It is suspected that press temperature is causing rejects in a plastic forming process. A Scatter Diagram shows a high positive correlation, prompting a redesign of the press, including the use of more heat-resistant materials. This results in a significant reduction in the number of rejected pieces.
• A personnel department plots salary against the results of a motivation survey. The result is a weak negative correlation. A second Scatter Diagram, plotting time in company against motivation, gives a higher correlation. A motivation program is targeted accordingly and results in a steady increase in the score given to motivation in the company's next personnel survey.

<-- Previous | Next -->