Thinking Critically about Data: Faulty data, fallacies and deceptions


Why do humans so often jump to faulty conclusions? How much is from bad data and how much is simply its misuse? In this course, you will hone your critical thinking processes to identify common statistical fallacies from the media, advertising and academic publications. You will learn to recognise common mistakes in the way data is collected, used and interpreted.

Recommended reading:
The course refers extensively to examples from the following two books, which we recommend you read:

  • Daniel Levitin: A Field Guide to Lies: Critical Thinking in the Information Age (Penguin, 2016).
  • Gary Smith: Standard Deviations: Flawed Assumptions, Tortured Data and Other Ways to Lie with Statistics (Duckworth, 2014).

Target audience:
This course is for anyone who is interested in the way statistical information is used: in the media, in academic reports and papers, and in advertising. It is aimed at a semi-popular level, with most examples involving simple arithmetic and graphs.

Learning objectives:
By the end of this course, you will have:

  • gained an awareness of common biases in human reasoning that lead to the misinterpretation of numerical information
  • learned to recognise common mistakes that arise from the way numerical data is collected, used and interpreted.

Course outline:
Different topics are covered during each session, with room for variation, depending on the interests of the group.

Session 1 looks at common and simple ways to misuse numerical information, noting the importance of context. Examples are taken from the media and published research.

Session 2 examines why untrained (and even trained) users of numerical information often misinterpret numerical and other data. This session considers why humans frequently jump to faulty conclusions.

Session 3 considers how data is generated. It covers the importance of context, garbage in, garbage out, and comparing apples with apples.

Session 4 highlights the importance of graphing data – you might be surprised at what you see. Learn how not to wittingly or unwittingly draw graphs that deceive.

Session 5 asks what causes cholera – bad air or bad water? Examine why we often blame A when, in reality, B is the culprit.

Session 6 looks at regression fallacies. The students with the best school results will likely do well but might not be those who do best at university. The weakest students may, by contrast, move up in the ranking.

A short break is held halfway through each session, and you are welcome to bring refreshments if you wish.

John Maindonald is a former lecturer at the Australian National University and a quantitative problem solver. He is the author of a book on statistical computation, and the senior author of Data Analysis and Graphics Using R: An Example-Based Approach (Cambridge University Press, 3rd edition, 2010).

For further information:
Continuing Education, Victoria University of Wellington, PO Box 600, Wellington 6140
Phone 04 463 6556, Email:

Please note: Courses need a minimum number of enrolments to go ahead. If your course doesn’t reach the number required, we’ll have to cancel it. If this happens, we’ll contact you by phone or email before the scheduled start date and arrange a full refund. Please check your emails regularly.