Harvey Goldstein on educating the public in using and critiquing official data

Harvey Goldstein of Bristol Uni, a legend in the world of promoting good statistical practice for league tables (by my own experience and others’, a small and demotivating world) , has written a blog post through the LSE on whose role it is to inform the public on how to access, use  and critique the great mass of data published by governmental and official agencies. There clearly is a problem here because hardly anyone goes to look at the data (as opposed to the often poorly digested versions in the media), unless they already have a degree in stats, social sciences, vel sim.

Prof Goldstein lays the responsibility firmly at the door of “academics” and also “the profession”, by which I think he means statisticians. So that’s me and a few thousand others around the UK, although blogs have an international reach because a lot of the data released is very similar.

I’m not sure we can do it. We’ve been banging on about it for years, both at the civil servants (who listen politely and then do what they were going to in the first place, but with the imprimatur of having consulted experts) and the public (who can read Nate Silver and Tim Harford books every Christmas and still have no idea where to begin in deciding which of their local schools is best). It needs to be a concerted effort at organisational level. The RSS and other such bodies around the world don’t generally have enough influence or resources to make it happen straight away, but they do get listened to by official agencies, and if we were more outspoken in criticising unhelpful or obfuscatory data releases, we would start to get some traction. Nobody in Whitehall wants their boss to print out the critical press release and slip it into the file marked “performance review”.

Farm rents graph from DEFRA. What are the error bars? How many respondents are there in each category? Will it print out OK in black and white? Why are there no error bars on any of the other graphs?

Farm rents graph from DEFRA. What are the error bars? How many respondents are there in each category? Will it print out OK in black and white? Why are there no error bars on any of the other graphs? Why not put it into a map?

Data should be published along with guidance and interpretation. It should be available for consideration at a number of levels, probably through some interactive website. It should be tabular and graphical and textual. That’s what we’ve been doing for years, and that’s what the really good official data producers are doing already. If we see powerful agencies not meeting those same standards of communication, we need to speak out and make a fuss. But there is no use in doing it in isolation.

Advertisements

Leave a comment

Filed under Uncategorized

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s