by George Taniwaki

In a paper authored by Erin Cooley, et al. published in J. Exper. Psy.: General, Apr 2019 (subscription required), there is a pair of line charts, reproduced below. I found them hard to interpret.

WhitePrivilegeChart_thumb1

Figure 1. Eight data points spread across two charts

Explanation of data

As I explain in a May 2019 blog post, the data is a result of two studies with eight sample cells, each represented by one of the points in the charts above. There are three pairs of groupings,

  1. Social group – social conservatives and social liberals
  2. Treatment – those receiving a lesson on white privilege and those who did not
  3. Race – those reading a story of a poor black man and those reading a story of a poor white man.

The table below summarizes the data from the charts.

Social group Race of poor person in story Received white privilege lesson Sympathy score
Conservative Black No 60
Conservative Black Yes 61
Conservative White No 53
Conservative White Yes 59
Liberal Black No 72
Liberal Black Yes 76
Liberal White No 71
Liberal White Yes 60

 x

The improved chart

Using the data from the above table, I created my own version of the chart.

image_thumb5

Figure 2. Rearranging the data from Figure 1

Let’s compare the charts.

In Figure 1, the data for of the groups receiving the treatment (reading a lesson on white privilege) and those who did not are plotted on separate charts. This make it difficult to track the effect.

Also in Figure 1, the horizontal axis is social group category. There are lines connecting social conservatives to liberals, two unrelated groups. Bar charts might be better than line charts to display the data.

In Figure 2, all eight data points fit on a single chart, making comparisons easier. Also, the point of this research was to determine the effect of providing a lesson on white privilege on sympathy scores. Thus, the horizontal axis should be the treatment categories, those who received the lesson on white privilege and those who did not. A line may connect the cases to show the slope, positive to indicate the lesson increases sympathy (desirable) or negative to indicate the lesson decreases sympathy (undesirable).

Finally, my chart has a smaller scale (50 to 80 instead of 50 to 100), fewer rule lines, and no box around the legend. All reduce clutter so you can concentrate on the data.