How to Read Numbers: A Guide to Statistics in the News (and Knowing When to Trust Them) - Kindle Edition

£0.99

Shared by

EnaHia

Joined in 2021

1,883

515

About this deal

Every day, most of us will read or watch something in the news that is based on statistics in some way. Sometimes it'll be obvious - 'X people develop cancer every year' - and sometimes less obvious - 'How smartphones destroyed a generation'. Statistics are an immensely powerful tool for understanding the world, but in the wrong hands they can be dangerous.

Introducing you to the common mistakes that journalists make and the tricks they sometimes deploy, HOW TO READ NUMBERS is a vital guide that will help you understand when and how to trust the numbers in the news - and, just as importantly, when not to.

More details at

Edited by a community support team member, 4 hours ago

New Comment

10 Comments

sorted by

3 h, 39 m ago
There's a simple answer when it comes to when to trust statistics in general and that is "never". Having spent most of my adult life (mis)using them - it's like card tricks: even when how it's done is explained most will still be mis-directed and not spot the truth. That was conscious bias. Unconscious bias and good old prejudice are always present. In the news, any statistic that doesn't support the publication's view of the World simply gets left out. Any that does, gets published no matter how misleading.
3 h, 16 m ago
Statistics can tell us a lot if interpreted correctly. The biggest issue is that people glady accept them without looking deeper if what the stats suggest matches how they want it to be.
3 h, 17 m ago
Couldn't agree more. I've spoken to people on opposing sides of an argument, all using the same statistics to try and prove that their side is right. In the end, most people just try and use statistics to support their argument, rather than intelligently form it in the first place.
1 h, 44 m ago
Stats are meaningless unless you also understand methodology and context.
1 h, 29 m ago
Problem is, that we all know the methodology and the context. The context is to support the argument - otherwise they wouldn't have been produced. The method is first to choose a dataset, or the part of one, that is most likely to support the argument. Then use mathematical trickery, eg limited precision numbers and ill-conditional matrices, so as to dilute and remove data from that dataset, that brings the argument into doubt. Then present the statistics in a way that creates an impression of supporting the argument, when it actually does not. eg offset zeroes, pictorial images with non-linear representation, colours that divert attention, etc.
Or, of course, just invent the dataset, knowing that most will just accept the result - if it correlates with their view of the argument. That works 99.9% of the time.