Monday, 18 June 2007

Shine a little light

Ages ago I happened upon this article on Wired about, a new(ish) development in mashups which aims to highlight links between campaign contributions and the votes that legislators cast once they have been elected. Maplight started out covering the California Legislature and promised in the Wired article to expand and track Congress too. This has now been implemented and my brilliantly planned tardiness can now be seen as a cunning strategy to wait until this project had reached its second stage before posting about it (ahem).

Digging into the data suggests some intriguing results, but is there really a direct correlation between the amount of money spent and the way in which a vote was cast? If there is, then doesn't an analysis of some of the votes suggest that a better (more targetted) use of funds could have produced a different result? This is obscured by the way in which information relating to a specific campaign can be presented as it suggests that the funds referred to have been donated solely in support of that particular measure. As this is not the case then we need to consider how much influence each donated dollar has across all the campaigns in which it might be relevant.

As a purely random example consider the result of a Californian measure dealing with forestry resources which failed to become law by 36 votes to 33. Supporters of the Bill donated around $105k, while its opponents donated about $267k. This suggests that the trigger value for a Yes vote was much lower. The Bill's supporters donated a total of $18.25k to people who voted against it (including $12k to one individual). At a 'price' of $2,648 per succesful vote this money could be said to be worth six extra votes (and a 39-30 victory). The point is that campaign donations are part of a much more complex web of influences and that generally (absent outright corruption) the value of each dollar in determining how someone votes is variable depending on a whole range of other factors, such as their moral and social incentives and other donations that are competing for their attention on this particular issue. Despite this I'm sure that there are some interesting patterns to be discovered here which can hopefully lead to difficult questions for elected representatives. At a time when access to information about our legislators seems to be going backwards again perhaps Maplight is also a good model of the kind of connected thinking that would be welcome in any new analysis of Parliamentary activity. Of course, this information could very useful not just to scrutinisers of accountability, but also to lobbyists. I think Henry Ford is supposed to have said that he knew that only 10% of his advertising spend was effective, but he had no way of knowing exactly what comprised the ineffective 90%. Ok, this is probably apocryphal, but Maplight's figures suggest that if donations were made with to intent to secure influence then some of them have failed. This project may make it clearer just what was wasted.

Maplight stress that their data only shows correlation, it does not prove causation. For that we have to rely on cartoons.

All of which leaves only one question remaining - how can Web 2.0 track brown envelopes stuffed with used notes?

Neatly tying in with some of the points that I was trying to make about the nefarious range of influences on legislators and politicians in general, Lawrence Lessig has announced that he is going to dedicate the next decade to the issue of the "corruption" of the political process.

No comments: