Lindsay Pierce/The Daily Times
Obama supporters surround McCain supporter Jim Krass at an Obama rally in Farmington, N.M., on Friday.
Obama supporters surround McCain supporter Jim Krass at an Obama rally in Farmington, N.M., on Friday. Lindsay Pierce/The Daily Times
Every election season, reporters fan out to states and counties that claim to be political bellwethers. After all, if the voters in these places have been right in the past, maybe they'll be right again.
But in presidential politics, there are actually few true bellwethers left. Missouri is the only state that has picked the winning presidential candidate consistently over time — in every presidential election since 1904, with the exception of 1956 (choosing Adlai Stevenson over Dwight Eisenhower, the winner in a national landslide, by less than 1 percent of the vote).
But consistently picking the national winner fits just one definition of bellwether. Two academics who studied the notion came up with three types of bellwethers. Princeton University professor Edward Tufte and his student Richard Sun defined electoral bellwethers this way:
The All-Or-Nothing Bellwether: These are states or counties that choose the national winner every time. Missouri fits this definition, as do six counties that have matched the national vote in every presidential election since 1960:
1. Vigo County, Ind.
2. Lincoln County, Mo.
3. Van Buren County, Ark.
4. Logan County, Ark.
5. Eddy County, N.M.
6. Ferry County, Wash.
The Barometric Bellwether: This is a place that accurately reflects the national share of votes. Vigo County, Ind., best fits the definition of a "barometric" bellwether.
The county has actually been correct in every election since 1892, with two exceptions (1908 and 1952), and has closely matched the actual national margin in every election since 1960. In those 12 elections, the voters of Vigo County have been within 3 percent of the national margins between presidential candidates. No other county comes close.
The Swingometric Bellwether: This is a county that mirrors important swings or shifts in the national electorate. Sandoval County, N.M., and Washoe County, Nev., seem to be examples of "swingometric" bellwethers.
Most of Sandoval County's population lives in booming suburban subdivisions. Some political scientists believe that these are the kinds of places that increasingly swing elections, because their voters are moderate and their political allegiances are soft. In fact, as NPR's Ted Robbins reports, "Four years ago, President Bush won New Mexico with fewer than 6,000 votes. That was the closest margin of any state, and the closest county in New Mexico was Sandoval County."
The race also was tight in Nevada in 2004. President Bush won the state by just 2.5 percent. And it has been one of the nation's fastest-growing states for more than a decade. That makes places like Washoe County (home to Reno) ripe for political gain, so the campaigns of both Barack Obama and John McCain see it as a key swing county in a key swing state.
For a complete list of bellwether counties and states and their voting records, try Dave Leip's Atlas of U.S. Presidential Elections.
Are Bellwethers Predictive?
Tufte and Sun actually studied the bellwether phenomenon by reviewing election returns for almost every U.S. county during the 14 presidential elections between 1916 and 1968. They wanted to know whether being right over time meant that bellwether counties and states could predict elections. And they were skeptics from the start.
"Prior to the 1936 presidential election, the conventional political wisdom had it that as Maine voted, so went the rest of the nation," Tufte and Sun wrote in a study titled "Are There Bellwether Electoral Districts?" that was published in Public Opinion Quarterly. They then noted the 1936 landslide, in which Maine and Vermont were the only two states NOT voting for the national winner, Franklin D. Roosevelt. You can almost hear the academics chuckling as they add, "Roosevelt's campaign manager revised the theory: 'As goes Maine, so goes Vermont.'"
Their point is that bellwether status doesn't seem to last. Every election, one bellwether or another falls off the list. Since the elections of 2000 and 2004, one state (Delaware) and two counties (New Castle County, Del., and Socorro County, N.M.) joined the former bellwether fraternity.
Tufte and Sun's 18-page statistical analysis is a demographer's dream and a number-phobe's nightmare. But they sum up their work in a single word. Are bellwethers predictive, they ask? "No," they answer. They suggest a random selection of counties may be better predictors statistically. The study closes with this zinger: "It is a waste of time to send reporters out to interview nonrandomly selected citizens of Crook County [Oregon] a week or two before the election — at least from any sort of scientific point of view."
Crook County had voted with the national presidential winner in 27 consecutive presidential elections. The voters of Crook County had been right for a century — from the time their county came into existence until 1992, when the county went for incumbent President George H.W. Bush while the nation chose Arkansas Gov. Bill Clinton.
In other words, bellwethers are like the stock market. Past returns do not guarantee future results.
And A Note About The Word 'Bellwether'...
You might be wondering where the word "bellwether" comes from. Just think about Mary and her little lamb. Its fleece was white as snow and everywhere that Mary went that lamb was sure to go. But if Mary had to visit the orthodontist, who did not tolerate sheep in the waiting room, she'd tie a bell around the neck of a wether (a castrated male sheep) who would lead the little lamb and the rest of the flock around until Mary came back. And when she returned, the bell signaled the flock's location.
So bellwether sheep are a lot like bellwether voters. We can tell where they've been (sheep leave tell-tale signs), and we know where they are, but we really can't be sure where they're going.