Scattergram

A scattergram is a visual summary demonstrating the relation between two variables; it illustrates the scores on one variable plotted against scores on another variable. The scattergram is usually drawn before working out a linear correlation coefficient or fitting a regression line. It is used as an aid to interpret the data. For instance, the scattergram can illustrate whether there are any outliers or gaps in the data or whether the data follow a straight or a curved line.

The pattern obtained from the scattergram illustrates the type and strength of the relation between two variables. If the data points make an upward trend going from bottom left to top right (positive slope), then the association is positive (the relation is direct; as one variable increases, so does the other). If the straight line goes down from top left to bottom right (negative slope), then the association is negative (the relation is indirect; as one variable increases, the other decreases). The closer the points follow a straight line, the higher the correlation. No visible trend from the scattergram implies no association; the points are scattered randomly. If the data points form a curved line, then a linear correlation cannot be used to describe the data.

In Figure 1, the scattergram illustrates a negative relation between age and traffic tickets. The scatter-gram was created using Excel.

When creating a scattergram, observe the following conventions:

1.
There are no gridlines for the X or Y axis.
2.
The axes are clearly labeled.
3.
The points are not connected by a line.

Figure 1 Total Number of Traffic Tickets Issued to Individuals Age 20 to 50 in 2005