Why are so many news stories based on surveys?
Because they are cheap to produce, easy to write up, and generate coverage reliably — and a substantial proportion are commissioned by organisations with a commercial interest in the result.
The mechanism. A public relations agency commissions a survey on behalf of a client, designs questions likely to produce a usable headline, and distributes the results as a press release with a ready-made story, quotations and sometimes images. A newsroom under time pressure publishes it largely unchanged.
Why these surveys are frequently worthless:
The sample. Frequently an online panel of people paid to answer questions, which is not representative. Panel members differ systematically from the general population, and response rates are rarely stated.
Leading questions. A question can be designed to produce almost any answer — through wording, ordering, or offering options that force a particular framing.
No methodology published. This is the clearest warning sign. A legitimate poll states sample size, method, dates, weighting and the exact question wording. A PR survey usually states none of it.
Selective reporting. Questions producing uninteresting answers are not mentioned.
Spurious precision. "73% of people say..." implies an accuracy the method cannot support.
Meaningless quantification — surveys producing a "formula" for something, or a national ranking of towns, are almost invariably promotional exercises for whoever paid.
How to assess one:
Who paid for it? This is the first question, and legitimate reporting states it.
Who conducted it? Membership of a professional body such as the British Polling Council brings disclosure obligations, including publishing tables.
What exactly was asked? The wording frequently does not support the headline.
How was the sample drawn, and how large?
Is there a link to the data? Its absence is informative.
Genuine polling is different and valuable — properly conducted, transparently reported, with stated margins of error. The problem is that the two are presented identically, and the same word covers both.
Some outlets now decline to cover commissioned surveys, and press regulators have received complaints about the practice.