THREAD: A thread on research and sampling, and how the media use data and polling.

I can safely say from a decade of working with media, it is *highly* unusual for a newspaper to so heavily reference a self-selecting and homemade survey and present it as the views of all people.

Surveys with self-selecting samples, in this case, both from FairFuelUK + cycling advocates, should not be represented as the views of all of those groups; same goes for claiming data represents "Tory voters".

And that's without even mentioning the survey's leading questions.

Polls can make for really interesting stories and are commonplace in media. YouGov, a key player, vets the questions and ensures they don't lead, and gets answers from a statistically representative sample to reflect the views of all GB adults.

You've probably seen the stories: X% of Brits think X. The sample size is over 2,000 and the methodology means they'd be confident the output would be the same, even if the number of people surveyed increased.

They ensure each demographic is represented as per the population.

I've worked with YouGov to conduct polls like this, most recently for @BikeIsBestHQ. Our nationally representative survey looked like this.

- 62% used their cars as first choice for errands
- Just 5% cycled
- 78% agreed the environment would be better if more people cycled