Monday, February 20, 2006

How good is an exit poll?

January 25th marked the second time that the Palestinians held parliamentary elections: according the BBC - "Exit polls suggest that the ruling Fatah party has won a narrow victory in the first Palestinian parliamentary elections for a decade. The first official exit poll suggests Fatah took more than 46%, compared to 39.5% for Islamic militant group Hamas. The official results may not be announced for several more days" Well official results did come out and Hamas won a landslide! getting 70% of the seats in the new parliament. How did this come about? Well, for about a month I have been saying "look! Exits polls are stupid since people can lie and will lie." Of course, this statement is often an accurate reflection of what is going on. But the numbers here tell a much more interesting story. According to final tallys Fatah won 41% and Hamas 45%. Yes, the numbers to reverse, but not so bizzarely as what the direct comparison between the exit poll percentage of the popular vote and the acutal percentage of awarded seats indicate. It turns out that the subtlety lies in the style of the democracy. What Palestine has is some combination of the American two-party system and the European/Israeli multi-party system. In the two party system a party can just barely win the popular vote and still win a landslide. That's what happened. . Dean can tell you more about why this is actually a very good thing. Dean?


At 2:08 PM , Blogger Dean Foster said...

The Wikipedia has a nice article on both purlality voting systems and the median voter. So let me say a bit about why I like the american system. If you think of voters as being homogenious over space (which is a fairly good approximation in the US, but not in say Iraq) then assigning districts leads to a way of generating something close to random samples of the population. Hence you need close to 50% of the entire population before you can get representation.

The advantage of this is that anything that 80% of the people agree on is no longer worth of fighing over. Just let go of it if you aren't in the majority! This means that only issues that have at least 40% of the people on both sides can be a real issue. These are things that aren't obvious--hence they are worth voting on.

Thus, I believe that as a statistician, the american system helps identify statistically signficant positions and get them solved. The ones that don't have a strong majority--should be left to a coin toss anyway. This is pretty much what we get.

At 9:57 AM , Anonymous Anonymous said...

Having lived under the proportional system for a while, there are additional problems: you are not voting for an individual who would be personally responsible and accessible. Instead, you're voting for a "list" of semi-anonymous representatives. Repeating this for several terms, one gets a distrust of the government, excessive role of the party-line and alienation of the people from their "representatives".

What Fatah should have done in Palestine is to concentrate their representatives in particular geographic areas. This is a popular policy in the UK.


Post a Comment

Subscribe to Post Comments [Atom]

<< Home