6 thoughts on “On the Public Understanding of – and Public Engagement With – Statistics: Reflections on the OU Statistics Group Conference on “Visualisation and Presentation in Statistics””

  1. Your final sentence wished for a new-to-the-majority statistical graphs with a strong narrative rather than maths and stats. Is OECD’s new Better Life Index http://oe.cd/bli the sort of thing you are wishing for?

  2. @toby I hadn’t seen that OECD site – thanks for the link:-)

    The current breakdowns are interesting (e.g. http://www.oecdbetterlifeindex.org/countries/greece/ ) in the way that text is used to explain and contextualise the vbar charts and each country’s position within a chart (which also invite the use to hover over different bars to see the neighbouring countries or extreme countries on each metric).

    The Better Life index is very shiny, but I wonder: do users get much more out of it than blindly and wildly changing the values in the right hand panel?

    In his presentation at the OU stats conference ( https://blog.ouseful.info/2011/05/18/quick-summary-of-opening-session-of-visualisation-and-presentation-in-statistics/ ), Michael Blastland gave a very open account about some of the interactive visualisations published by the BBC (links in the above report), admitting that in many cases they just didn’t work. For sure they were shiny and offered open ended exploration possibilities, but in many cases the user arrived with no context that allowed them to play meaningfully, or in a directed fashion?

    Part of the point of these interactives is surely related to trying to help users develop or understand the model the interactive represents, and then test their understanding/appreciation of the model against the interactive?

    (I think I need to start reading philosophy of science books again!)

  3. I think the overall point about engagement (as opposed to understanding) is a good one, and we statisticians should think in those terms, but the situation in statistics is different from in science. Putting it far too crudely (and arguing against some things I said at VIPS!), statistics largely consists of methodologies for exploring questions that aren’t themselves really part of statistics, and so it’s even more important than in science to persuade people that engaging with statistics is a way of getting to things that they really want to know or do, and which might well not look very statistical at all. This links to what Michael Blastland was saying; one way of looking at that is that the shiny tools are no use without having got people interested in the questions you can answer with them. The distance between what the statistics can at face value tell you, and the things you actually want to know, can be pretty great — bridgeable, but it’s a long bridge often. (And here I’m talking about /any/ sort of statistical information or tools, really, including the new kinds of visualisations.) Other issues are that people already /do/ engage with statistics in some areas (e.g. sport) on a major scale, so there’s important stuff to build on in further engagement — that’s good, but brings its own challenges. And that the nature of “citizen statistics” looks (so far anyway) very different from “citizen science”. The science tends to be stuff that’s led and directed by professional scientists, and the citizens help with collecting data and making comments and looking for patterns in a directed way, and that sort of thing. “Citizen statistics” in the sense of using (and understanding) public data and public tools is different, in that, in principle (and sometimes in practice too) it really is done by individual “citizens” or groups of them, to deal with their own concerns and answer their own questions, and the input from the professional statistician or data scientist (or whatever one wants to call them) can be limited to help them use, appreciate, engage with the tolls and data sources rather than setting the agenda.

  4. @kevin I think making the distinction between engagement with science vs engagement with statistics is an interesting one and is something I will ponder. I’m also considering replacing the word “statistics” with “tool” or “technology” when developing my ideas/arguments in this area, and then as a sideline consider what it would take for “the public” to see stats in the same way. (As I think you hint at too, there may be a mismatch between how different audiences think of statistics and what statisticians do, which just adds clutter and confusion to the engagement exercise.)

    I never thought it would be stats that would rekindle my interest in the practice of science and technology engagement/communication!

Comments are closed.

%d bloggers like this: