Misleading Graphs - How Can We Spot Them?

In today’s technological age, graphs have been extremely helpful in showing information between a set of options. They can be seen everywhere: from toothpaste ads stating that 4 out of 5 doctors recommend their brand, to politicians claiming that they have decreased house poverty by 25% in a couple of years. Many people already know that when somebody simply states a claim like this, it is most likely exaggerated or overemphasized. However, when they see a graph with information on it, they believe in people’s statements more, even if the graph doesn’t show the entire story. So how can we figure out whether a table is deceiving or completely true?

Problem 1: Cherry Picking

One very popular way in manipulating what a graph states is by using a technique called “Cherry Picking”. You see, advertisers can’t simply change the graph’s information, so they have to find some workarounds. Cherry Picking in graphing terms is when people choose a certain point of time to maximize their graph’s looks. An example of this is a chart showing the increase of American job loss. When looked at first glance, the table showed a comically straight line, representing an “heavy increase” in unemployed Americans. However, once you take a look at the timeframes, you can immediately see a misleading point: the time distance between points is not even, with points being either 6 or 15 months apart. However, that is not the main influencer, as the normal graph with even time slots still shows an increase. The time frame chosen was directly after the Great Depression, a time with low employment. This example shows a great influencer in changing graphs: time.

Problem 2: Weird Scaling

When showing a scale, acknowledging how your values are presented can be a great factor. An example of this is when Chevy, a truck manufacturer, released an advertisement in 1992 apparently stating that they had the greatest reliability of all truck brands. To support their claim, they added a graph, showing that Chevy had 98% of their trucks still driving. In the way they presented their graph, Chevy seems more impressive than their competitors, with Honda, one of Chevy’s biggest rivals, only doing half as well. However, by looking at the percentages at the right, the truck company seems a lot less impressive. Honda, which apparently is supposed to be half as well as Chevy, has 96.5% of their cars on roads. Then you notice the big factor: the table only shows the percentages from 95% to 100%! If they set the percentages to 0% to 100%, the differences between the brands would be barely noticeable.

Problem 3: Background Information

Another major contender in graphing is knowing what it represents. One example of this is a graph of traumatic injuries inflicting children. Most of the labels there are mildly concerning, but the biggest, most scaring percentage shown was the one surrounding spinal cord injuries, being at “5.2%” for children. While many people, including both parents and children, could be terrified of this outcome, it is not as scary as it seems. When looking at where the percentages come from, you can see that the graph is based on 2,000 injured children a year out of 74,000,000. (Using real calculations, we can determine that the real percentage is “0.0000003%” for a spinal cord injury.) 

Conclusion

The effects of graphs have affected our description of information all over the social media world. However, there are many things we must keep in mind whenever judging these occasionally untrue graphs. The time, scaling, and background of a graph can be essential to what it shows, and can be edited to increase one’s attention. Knowing this, make sure to not immediately jump to conclusions every time you view a graph.