The Guardian

How do you know whether you can trust poll results? Here’s what to watch out for | Rob Vance

The industry adage is that if it looks too good to be true it’s probably wrong. But the data reveals what’s really going on
‘Treat polling findings in a critical and informed way, subject to the potential bias you look for in any writing.’ Photograph: Rancz Andrei/Alamy

With the science of opinion polling under greater scrutiny than ever, and a somewhat misleading headline spreading in an instant, how can you know which? Perhaps counter-intuitively, it’s less about sample size (let’s avoid a poor joke) than about quality. You will rarely get less than 1,000 interviews reported in the UK media, but twice the sample size does not mean double the quality. If there are issues with your sample then making it bigger doesn’t necessarily improve it.

You’re reading a preview, subscribe to read more.

More from The Guardian

The Guardian4 min read
‘Everyone Owns At Least One Pair’: $75bn Sneaker Industry Unboxed In Gold Coast Exhibition
What was the world’s first sneaker? Was it made in the 1830s, when the UK’s Liverpool Rubber Company fused canvas tops to rubber soles, creating beach footwear for the Victorian middle class? Or was it a few decades later, about 1870, with the invent
The Guardian4 min read
Lawn And Order: The Evergreen Appeal Of Grass-cutting In Video Games
Jessica used to come for tea on Tuesdays, and all she wanted to do was cut grass. Every week, we’d click The Legend of Zelda: The Wind Waker’s miniature disc into my GameCube and she’d ready her sword. Because she was a couple of years younger than m
The Guardian4 min read
‘Almost Like Election Night’: Behind The Scenes Of Spotify Wrapped
There’s a flurry of activities inside Spotify’s New York City’s offices in the Financial District. “It’s almost like election night,” Louisa Ferguson, Spotify’s global head of marketing experience says, referring to a bustling newsroom. At the same t

Related Books & Audiobooks