Skip to main content

Quick Tips On Writing with Statistics

OWL logo

Welcome to the Purdue OWL

This page is brought to you by the OWL at Purdue University. When printing this page, you must include the entire legal notice.

Copyright ©1995-2018 by The Writing Lab & The OWL at Purdue and Purdue University. All rights reserved. This material may not be published, reproduced, broadcast, rewritten, or redistributed without permission. Use of this site constitutes acceptance of our terms and conditions of fair use.

1. Never calculate or use a statistical procedure you don't fully understand. If you need a statistical procedure, and you don't understand it, then you need to consult someone or learn how to do it properly.

2. Never attempt to interpret the results of a statistical procedure you don't fully understand. If you need to interpret a particular statistic, talk with a professional statistician and make sure you understand the proper interpretation. Unlike descriptive statistics, inferential statistics is anything but black and white, there may be several valid interpretations of a given statistic, and you need to be aware of which ones are better under which circumstances.

3. If you are using statistics in a paper, consider your audience. Will they understand the statistics you are using? If not, you may need to explain the procedure that you are using in detail. This is not inappropriate! It is better to include too much information than too little. Depending on your field, this may be done using an appendix, footnotes, or directly in the text.

4. Present as much information as needed so that your reader can make their own interpretation of your data. Certainly, your job is to help the reader interpret your data, but most statistics are used to support a persuasive argument. You need to give your reader enough information that they can reconstruct your argument from your statistics. If you don't give them enough information, people will think that you are being deceptive, which can damage your credibility. You can't convince someone of anything if they are convinced that you are misleading them!

5. Use graphics and tables. Statistics can contain a lot of information, and using visuals can display a lot of information in a manner that can be quickly understood. See the section on visuals and statistics for more information.

6. If it's applicable, and you can calculate it, do include some measure of variability; typically this is a standard deviation. Even if you aren't doing any inferential statistics, this statistic provides excellent information about your data set.

7. Be wary of using statistics from other places that are not peer-reviewed. Popular magazines are notorious for including bad statistics. Oftentimes their 'sample' is a section of people who choose to respond to some online query. Their sample often includes mostly women or mostly men (depending on the magazine) but rarely do they have a good representation from all genders, and many times the magazines imply that the results generalize to the entire population. While some statistics might be generalizable, many are not. If it's not from a reliable source, then don't use it.

8. Speaking of sources, if you used a statistic, you need to provide your audience with additional information including where the statistic came from. You should be wary of statistics that seem to appear out of nowhere.

  • A poor example: The ten largest cities in the U.S. comprised 54% of the total U.S. population.
  • A good example: According to the United States Census Bureau, in 2000, the ten largest cities in the U.S. comprised 54% of the total U.S. population.

In the second example, your audience knows exactly where the statistic comes from (if they don't believe your statistic, they can go and check themselves) and it comes from a reputable source (the U.S. Census Bureau).

9. If you calculated a statistic, how did you calculate it? In some fields, you don't need to tell your readers how you calculated some statistics. For example, in psychology, you don't need to explain how you did an ANOVA or a t-test, but in other areas you might need to explain this in more detail.

10. Be clear as to what population(s) your statistic is meant to generalize to. If your sample included only male college students, you should be very careful if you want to generalize your results to female lawyers. Don't imply that your sample generalizes to everyone if your statistic was calculated from a more specific population.

11. If you are using inferential statistics, try to speak as plainly as possible, and put the statistics at the end of the sentence. See the Writing Inferential Statistics section for more information.