How You’re Incorrectly Using Statistics in WritingMarch 21st, 2018 by Sara Vazquez Shaw
Using statistics in research or any other type of writing can be a powerful way to convey a point. However, when statistics are used incorrectly or misleadingly, it can bring what you've written into question. Poor statistical writing runs the risk of spreading misinformation. In legal or medical writing, this type of abuse, whether intentional or not, can yield some dire consequences.
In order to avoid being the bearer of misinformation, consider these tips to help you use statistics more responsibly:
Consider the Source
Publications sometimes cite statistics without specifying the studies they originate from. Always consider the primary source: where did the information come from originally, is it reliable?
Understand What an Average is
People use the term "average" frequently in everyday life. As such, writers often assume that readers automatically understand what "average" means. To use averages in the most accurate way, you need to be more specific. Is it the average, the median, or the mode you're referring to?
The average is determined by taking all numbers in a group, adding them up, and dividing by the amount of numbers. The median is the number precisely in the middle when all numbers occurring in a set are put in order. Lastly, the mode is determined by identifying the number value that occurs most frequently in a series of numbers.
Use Statistical Terms Correctly
It is not so much the confusion of causation and correlation that cause writers a problem; it is the fact that they are often used incorrectly to justify a point. For example, if the rise in a city's crime rate correlates to a rise in the number of power outages that city experiences, it doesn't necessarily mean that the rise in crime rate is caused by the power outages. It's important that writers explore correlation and causation separately. Writers should not draw conclusions based on what could be coincidence or circumstances that are provoked by other factors.
Account for the Margin of Error
Statistics about groups of people are typically derived from surveys. Surveys are limited in that they do not account for an entire population, only a sampling. Therefore, there is always a margin of error that might change the survey results by a few percentage points. The margin of error percentage is commonly very small, but if there is no acknowledgement that it exists, the reader is not getting a fully accurate report of the numbers.