Every month we highlight some of the statistics and survey
results buried in press releases from the ResponseSource Press Release Wire. Polls
are a controversial source for stories: we’ve all read the small print on TV
cosmetics ads and realised that 82% of users who would recommend a moisturiser isn’t
quite so conclusive when the sample size is 38.
But a survey with a representative sample of a relevant audience,
with details about the methodology and background to the research, can form a
news story in itself as well as supporting a feature or providing an
interesting snippet for your social media. The stories we highlight from the
Press Release Wire might fall under “fun facts” as often as they do under
serious statistics, so we’ve pulled together some further links for evaluating surveys
Sample sizes When a brand quotes its own
consumer surveys, what makes a fair sample size and what conclusions can
they legitimately draw? The Advertising Standards Authority’s guidance on
substantiating surveys and sample claims is here.
The Poynter Institute offers a free self-directed course on understanding and
interpreting polls here
or if you prefer a UK perspective, try the Royal Statistical Society’s (RSS) free
Statistics for Journalists course here.
For even more depth, the European Journalism Centre’s Datajournalism.com offers
free video courses and two comprehensive handbooks here.
Industry associations In November 2019 IMPRESS and The Market Research Society (MRS) brought out
a handy 11-page guide “to provide journalists with the tools they require to
report on statistical data responsibly”. You can download it here.
The MRS and RSS also joined forces with the PR industry’s
CIPR to produce a best practice guide for public relations professionals here.
Polling The British Polling Council, which promotes professional standards in
public opinion polling and aims to increase understanding of how polls are
conducted, provides specific guidance for journalists here.
Journalist’s Resource has created a list of questions you
should ask about every poll here.
The FAQs from YouGov here give
plenty of background in to their panel and methodology – for other polling
organisations see the British Polling Council’s members list here.
Margins of error Journalist’s Resource has created a useful guide here
to help you understand the margin of error in stats and why you should be
cautious with results (and interpretations).
Media policy Major media outlets tend to have policies on surveys, polls and statistics. The most comprehensive we’ve seen is from the BBC here and more advice on reporting statistics here.
ResponseSource’s top three tips when working with stats:
– Determine the author of the report and interrogate their methodology. Is the sample representative of a population and is the sample size significant? A UK sample typically needs to be above 2,000 to be nationally representative. As well as the resources above, a useful overview can be found here.
– Do your own calculations. It can be worth asking to see
the raw data behind the research; not only can you satisfy yourself that the
original story is accurate but there may be more: Datajournalism.com’s Data Journalism Handbook
points out “Data as such is unwieldy…It needs experienced journalists, who have
the stamina to look at often confusing, often boring raw data and “see” the
hidden stories in there.”
– Even if you’re confident about your own expertise, an independent point of view can add to your authority. Industry organisations in any sector can often connect you with experts via their media centre and universities will normally provide full details about the experience and specialisms of their academic staff. If you’re short on time, our Journalist Enquiry Service community can connect you with thousands of sources business, finance, technology, retail, health and any other sector – it’s still up to you to evaluate what they say – send a request on https://responsesource.com/send