Confusing Statistics



When analyzing data, we usually using different calculation methods in order to perform the data in a more understandable way. However, some of the methods not only make the presentation much more difficult, but also confusing. Below, I will briefly introduce 3 kinds of confusing Statistics, and look into one of the confusing method, Percentage .


Percentage

It was easy to be confuse when using percentage to present data. Although percentage may let readers realizing the ratio and to get the whole picture. However, when comparing 2 groups of data, percentage is not a suitable comparative tools. It was because the sample size data may not be the same. If they only compare through percentage, it may show the wrong concept to readers. We will discuss it further in below.
Summation

Summation may help readers know more about the result of the data, but showing data without showing the numbers for each part, may cause misleading. When we doing summation, we need to consider that whether the data are from the same category. For example, if we want to say the US people are rich, and using the data showing the summation of all income of US people. However, in the real situation, 99% of wealth in US are owned by 1% of US people. Many US people are still suffer in poverty. However, the summation may hide the situation and mislead readers

Skewness

We need to be careful when choosing the suitable skewness. Median, mean and average was the most common skewness. In different situation, we have to choose different type. For example, when we are facing extreme data, like 0, it is better to use median or mean. Otherwise, the average cannot represent the common amount since the extreme data affect the average.

Percentage

When using percentage, we need to be careful. From the graph we can see the percentage of participated students in Laboratory. It was clearly shown that the percentage of year 2 students was much higher than year 3 students. Readers may get the concept that more year 2 student attended Laboratory than year 3 students, However we can see it was false in the next graph.






From the graph using the exact data, we can see that the actual numbers of year 3 students participating in Laboratory have never been less than year 2 students. In fact, more year 3 students participated in SPY than year 2 students. The result is totally different from above. The reason that caused this problem is wrong using of percentage. The sample size of year 3 students is 50 while only 30 year2 students. To avoid this problem, we need to pay more attention when using percentage.


Chart with exact data

Learn More