By stewartb

What an appalling report: I’ve now read it through twice and remain confused – even finding the questions asked is a challenge!! To help understand the claimed conclusion in the newspaper about 69% opposing (something!) I searched the text specifically for every occurrence of ‘69%’ – I am none the wiser.

Two others things strike me. Firstly, there seems to be use of ‘straw man’ propositions throughout. Their inclusion may confuse and/or ‘condition’ and therefore slant responses.

Secondly, (and candidly, this is as best as I can disentangle from this report – but I may well have been ‘confused’), the one set of questions which arguably gets closest to addressing the core issue is far from giving a clear cut message from respondents. I refer to Question 2.

It states: “Q2. Summary: If each of the following claimed that certain words were ‘abusive’ and ‘likely to stir up hatred’ against them, to what extent do you support or oppose someone who used those words being prosecuted for a hate crime?”

There are then seven distinct characteristics listed including but not limited to ‘a religious person’, ’ ‘a trans person’ and ‘an elderly person’.

For each of the seven named characteristics, the percentage AGREEING i.e. supporting prosecution for a hate crime varies between 27% and 40%; the equivalent figures for those OPPOSING, i.e. against prosecution for a hate crime varies between 28% and 30%.

The highest level of support for prosecution for a hate crime -when it is “claimed that certain words were ‘abusive’ and ‘likely to stir up hatred’ against them” – is when the act/s are aimed at disabled people (40%).

Is this yet another example of how survey methods are abused and their results used and abused? How many readers of the newspapers publishing such headlines will actually go back to source?

The use and abuse of survey results has been an issue raised a number times on this site. By chance, I came across this earlier today from the specialist online news source ‘Nursing in Practice’.

Its headline is : ‘Nursing in Practice survey: Care homes grappling with extra work and reduced staff’.

The article begins: “Workloads for social care staff across the UK have risen since the Covid-19 outbreak, according to a Nursing in Practice snapshot survey.”

The article then explains: “All 12 respondents in the survey, run in July and August, said their workload was either ‘higher’ (7 people or 70%) or ‘much higher’ (3 people or 30%) since the pandemic began. “

So 12 respondents contacted over two months constitute a survey! And such a survey enables a conclusion to be drawn about social care staff across the UK! Should this win a prize for the most unrepresentative, statistically bizarre survey ever?


It would be hilarious if this abuse of survey evidence was not a serious problem.