Spread Plot

 

Menu location: Graphics_Spread.

 

This is a useful way to show the spread of data across groups. It is one step back from the Box and Whisker plot in that it gives an entirely pictorial representation of the spread of your data. The scale axis is divided into an arbitrary number of divisions which are the width of a plot point; if more than one data point occupies a division it is plotted alongside the first, thus a concentration of data at a particular value is represented by a broad band.

 

Example

Test workbook (Graphics worksheet: SDI conceived, SDI not conceived).

 

The plot above shows the Sperm Deformity Index (SDI) values of Aziz et al. (1996), from semen samples of men in an infertility study, in two groups: the 42 men whose partners conceived and the 116 men whose partners did not. The same data are used in the ROC curve example.

 

To draw it in StatsDirect open the test workbook using the file open function of the file menu. Then select Spread from the Graphics menu. Select the columns marked "SDI conceived" and "SDI not conceived" when prompted for data. The first column selected is plotted at the top. The chart above is titled "Sperm Deformity Index and Conception"; otherwise the chart is titled "Spread plot".

 

For this example:

 

SDI conceived: n = 42, range 122 to 239, median 165.5

SDI not conceived: n = 116, range 107 to 193, median 140.5

 

SDI conceived: broadest band 6 values at 162

SDI not conceived: broadest band 15 values at 138

 

A band is a column of values that lie within the width of one plot point of the first (lowest) value in the column; on the scale from 100 to 240 of this plot a plot point is about two units of SDI wide. The men whose partners did not conceive form a dense band between about 125 and 160, whereas the values of the men whose partners conceived are spread more widely and lie mostly higher. The spread plot shows the two distributions without summarising them; the ROC curve example measures how well the index separates the two groups.