The UK Statistics Authority’s event on “Better Statistics, Better Decisions”, including a session on what the public wants from statistics. I wasn’t there but the organisation, The Democratic Society, has a write-up.
My wish list is below. But first, a comment about “the public”.
Some years ago I gave a talk at the annual conference of the Royal Statistical Society. I started by asking members of the audience to put a hand up if they were from, say, the Office for National Statistics, or a particular government department.
I then asked: “Are there any members of the public here?” Half the audience put their hands up, including the then National Statistician, (now Dame) Karen Dunnell.
My point is that we are all members of the public.
As soon as we start thinking of the remainder of the public as a separate body of people – and of ourselves as some expert group – we magnify the chance that what we do will be affected by our biases, prejudices and preconceptions about the interests, abilities and capacities of the others.
And now my wish list…
I want to be able to go to official sources and find the key numbers right there in front of me – no need for me to read lots of text, hunt through tables, try to make sense of acronyms and codes. Of course, I also want to be able to get the data or read the ‘small print’ from time to time. Good use of the web can achieve this.
This is what I would like in detail:
- a number or a range, including an appropriate expression in words (particularly if the number is a percentage)
- the source organisation
- the exact source publication, and where within it, eg including row and column of data if necessary
- the release date and the date or time period to which it relates
- the next release date if known so a calendar reminder can be set if wanted
- any interpretations made
- any assumptions made
- the level of uncertainty and the degree of confidence
I’d also like to see links to other useful resources, such as visualisations. For visualisations, I’d like a short description of why the data is presented in that way. For example, if the y-axis doesn’t start at zero (sometimes called broken or interrupted) then why. Highlighting variation might be good, but sometimes it’s better to show the relative lack of variation, for example.
So, that’s my list. What would be on yours? Tell the UK Statistics Authority!
UPDATE 1 (25 November 2014): and an end to pre-release access – I know that’s down to the politicians to end but I can hope! Thanks to @statsgeekclare on Twitter for reminding me of this.
UPDATE 2 (25 November 2014): this has had some feedback from around the world. Although initially intended as a list for the UK Statistics Authority and so for UK official statistics, I can see it could be good for any statistics producer. Here’s an additional suggestion for the list from Twitter’s @SigmundChuck via @DiegoKuonen:
- what motivated the work/research, how it was funded