Friday, June 29, 2012

A synthesis

Today, I'm going to attempt to draw on a variety of different pieces of writing that I've read in the past few weeks. The end result, is, I'll admit, rather utopian and thus I'm probably dreaming if I ever think this will come to pass. Still, at the least, it's something worth working towards.
The journey starts with a peice by a chap going by the name of Tom Steinberg, in which he implores governments, regional and national, to stop treating people like idiots. The basic theory goes that governments are still in paper mode - in that sharing of information is difficult and limited, when in fact there is a huge opportunity to share not just the decisions but the rationale behind decisions. So a council might put something on it's website about rubbish collection being fortnightly now rather than weekly. And miss the opportunity to get a random member of the public more involved/informed by linking to the reasons for the decision, explaining the costs involved, linking further on to a budget sheet showing where the rubbish collection sits in the greater scheme of the council finances. It's an idea that makes enourmous amounts of sense to me. People are not dumb (a large majority of them aren't anyway) and you might have less people pissed off with the way things are running if it's explained in context. Numbers often get bandied about in the news, but rarely do I see them compared. Which can lead to people not really understanding how everything fits together.

So people have to have access to data - which brings us to the open government project, which I talked about a few days ago. There is a wealth of data that our various governments generate. It's good to have a central point which makes access easier - at the same time it's rather disappointing that we have to. The open government data is a goldmine for data driven journalists (the best of which that we have being Keith Ng atm) and for that reason alone it should exist. A lot of this information should incorporated into the fabric of government though. You shouldn't have to go searching through an archive to figure out how much money is being spent on anti-smoking measures this year. It should be linked in from district health boards alongside information on how much each smoker costs, how much getting them to stop costs and how much that is going to save us in the long term. With the option of clicking through somewhere to see how anti-smoking measures compare with other preventative programs in place.

Which brings me to a piece by Ben Goldacre. Or rather a report for the UK cabinet office that he co-wrote on randomised controlled trials. Through the use of numerous examples and some very plain language, they explain why randomised trials are good - you can find out what actually works and how well it works rather than just guessing or relying on biased analysis. They explain how to go about setting up randomised, controlled trials. They show why they don't always cost a huge amount more - the data is often already being collected. And get around the old chestnut of it being unetthical to offer something you will improve outcomes to one bunch of people and not another - the thing that you think might be improving outcomes could be hurting outcomes and RCT's can often be done as part of a staggered rollout, amongst other things. They suggest using the expertise of academics who know how to do the stats, which is a bonus, academics become more invovled in the policy process without taking it over and possibly get a paper or two out of it, properly managed, I don't really see a downside to this.

This then is where I get all utopian and dreamy eyed. Imagine a government that integrated it's data, made it people to access and attached it to relevant information on the web - so that you could easily go to whatever level you liked to see where it fit in with the rest of what the government is spending/doing. Not only that though, but instead of rocking on into government, loudly announcing that they going to implement their policy X, that policy X was going to fix everything and just steam rolled through, imagine a government that came into power with a bunch of trials they wanted to run over varying time frames and that based on the outcomes, those policies would be implemented. And to top it off, there's a decent chance that it would result in continually better outcomes such that the world of 2032 would be a better place for everyone than 2012 is.
I don't think this is necessarily the method to electoral death. If a party could build itself a reputation for building polices with properly collected data to back them up they could portray the other parties as ideologues with no interest in improving the lot of the average citizen. A position that I would like to think would at least win some respect.

No comments:

Post a Comment