Saturday, March 24, 2007

Figures Don't Lie, but Liars ...


... make maps.

Look at the map above. A similar one showed up in the pro-viaduct Seattle Times this week. The actual vote used to produce these maps said 70% no to a proposed waterfront tunnel, and 57% no to a viaduct rebuild. But the WAY we present the information distorts these facts in three interesting ways:

1) The Brand X Comparison: They compare the tunnel to the viaduct. The viaduct was only slightly less odious than the tunnel, so therefore the maps show more people preferring the viaduct. If you look at the map in passing, you'd assume that the viaduct won, when in fact both lost. The two, oddly enough, are not connected, and lack of preference for one does not indicate a preference for the other.

2) Land Votes, Not People: The denser urban core (which is going to have to deal with both the inconvenience of construction and the final product) voted strongly against the viaduct. But that is a smaller space than the more expansive areas in the west or north which had fewer voters, but occupied more territory. This is the scam used by the GOP whenever they do a national map, where Montana is a huge swatch of red, while the more populous NYC is an ignorable dot. Its a good way of thinking if we're thinking about 17th Century Virginia, where only landholders got to vote, but less accurate here.

3) Rounding Up: Looking at these maps, you'd assume that the majority likes the viaduct idea. After all, a lot of the area is colored in. But only the deepest shade makes any real difference, as that is the 50% plus category. In every other location other than the deepest red, this vote lost. By claiming everything of 30% up as a win, the viaduct is claiming support where it actually has none.

The three items, taken together, produces as fraudulent an image as Reagan's tear on the recent cover of TIME. Then, who are you going to believe, the facts, or your lying eyes?

More later,