There is an old saying: "There are lies, and then there are damn lies." Last week's post quoted Machiavelli, writing, "occasionally words must serve to veil the facts." If presented with artful deception, charts and graphs can also veil the facts. Recently I found an interesting paper on the ways statistics and other information can be presented in such a way as to mislead, while remaining arguably accurate. "Five Ways Writers Use Misleading Graphs to Manipulate You," Ryan McCready, September 11, 2018. .
The graphs I am using in this blog come from Mr. McCready's paper, and I strongly recommend that you go to his site to explore his full paper. So much information is presented to us in charts and graphs, and we need to be able to recognize when it has been presented in a intentional way to mislead.
The two bar graphs above present the same statistics, but can you see what has been done to manipulate your reaction to the numbers? Which one gives you the more accurate impression of the information? Notice where the bottom of the graph begins. The one on the left begins at 50, and using that manipulation of the numbers, it appears to exaggerate the differences between Group A and the other two groups. As Machiavelli would have said, 'the facts have been veiled.'
Again, the same statistics are presented in these two graphs, but notice the numbers on the left side on the charts. The chart on the left goes from 0 to 40, producing a flatter climb, while the chart on the right goes from 0 to 15, creating an abrupt climb between 2019 to 2020.
This time the trickery appears at the bottom of the charts shown above. The chart on the left shows monthly changes, while the chart on the right shows yearly changes. If the text or title make clear what the chart represents, there might be no trickery, but if the user has cherry picked data only from the period that reinforces his argument, the data can mislead.
Author Ryan McCready has included this chart to show that to convey accurate or easily read information, one type of chart may be better than another. He suggests that bar graphs are better for showing differences between groups. In the graph on the right, Team B stands out, while Team C lags behind both other teams. Yet, if you look at the round graph on the left, the differences are less apparent.
McCready also points out that sometimes using different colors can confuse, especially if one color is commonly used to represent particular information. Thanks to Ryan McCready's paper, examples of how abuses through the presentation of information may be as deceptive as words can be, in other words, examples of ways to 'lie without lying'. (See 11- 6 -2019 blog below.)
Battling Abuses by the manner in which graphs present information may be more difficult for muckrakers to expose, but I hope this week's blog has made you more wary about accepting information contained in charts and graphs. Next week the Series concludes, and if you have not been reading the full series, you may wish to scroll through earlier posts.