Latest Blogs
Blog image
Bad Graphs

There are a lot of people out there with bad opinions. We all think differently, believe different things, and process information in our own way, so the combination of all of those factors can lead to differing ideas and thoughts that you might believe are insane. When someone tells you something that seems too good to be true, your first thought is to ask them to prove it. That is fairly standard as the minimum thing that you need to back up a thought. If you can show your work, it gives validity to the things that you say. With that, people like to use facts, figures, data, and charts to back up their ideas. This is a great way to communicate information, but it can also be used by the forces of evil to corrupt the information that we receive. Media literacy matters and the ability to discern good from bad data is a skill that we should all possess. Here are a few ways to look out for how people will manipulate graphs to manipulate you.

Sample Size

In any sort of study, it is important to know the size of the sample. In any case, when you are looking to gather data, the larger the sample that you can create, the more accurate data that you will receive. For example, if I were to tests the idea that a coin flip is truly 50-50, I might flip a coin over and over and record the results. If I only do this two times, there is a 50 percent chance that I get the same thing both times. When you increase that to 5 tries, there is only a 6% chance of hitting the same thing every time. As you increase the sample size, you are much more likely to flip a similar number of heads as tails because the numbers normalize over time. Sample size is not always as simple as just collecting data point because you might be limited in scope or money. That said, a small sample can’t be seen as representative, no matter how hard you try.

Inconsistent Intervals

This might be the most prevalent and dangerous, so it is important to always look at the two axes when looking at any kind of bar or line graph. An example of this going wrong is when the time frames on the x-axis are not equal. If you start with a three month chunk for the first data point, you can’t use a 10 month chunk for the second. When you start to mix time frames, you are comparing what you want to compare, not the objective facts. You might also notice an oddly narrow graph to highlight a large difference. If you start the y-axis at 32 and end at 38, you will notice a stark difference between 33 and 37, especially if the increments are by halves or fifths. If you put that same data in a chart where the y-axis is going from 0 to 100, they will seem relatively close. It is all how you choose to present the data and that can be used for both good and evil.