Scatter Plot

Go Back

Definition

A scatter plot or scatter diagram is a graph generated by plotting paired observations on continuous variables. One variable is represented along the X-axis while the other takes values along the Y-axis.

Once enough points are placed on the graph, a general trend usually emerges, showing how the two variables vary together. If one increases along with the other, there is a positive correlation between the two. If one decreases as the other increases, the correlation is negative, or inverse.

Although one may treat the X-variable as a predictor and the Y-variable as the response, a scatter plot does not by itself make these distinctions and hence a conclusion of causal relationship between the two plotted variables must always be based on knowledge from other sources/analyses of such a relation between the two.

Examples

The scatter plot in this example shows the amount of sales per month generated at two locations of a chain grocery mart for the past year. The X-scale shows the sales (in thousand dollars) for Location A while the Y-scale represents the corresponding sales for Location B. The plotted points form a negative slope - the sales at location B is inversely related to the sales at location A.

Does this mean that location A caused the decrease in sales at location B, or vice versa – that the decrease in location B’s sales caused an increase in sales at location A? Not necessarily - unless the two locations were direct competitors, this phenomenon could be explained by several other equally valid other reasons, such as: the regional population at location A increased while that at location B decreased, or a change in management at location B caused the quality levels there to plummet, resulting in lost clientele, etc.

Correlation Analysis
Correlation Coefficient