Monday, October 25, 2004

Best signs for predicting the winner

Question: Which polling organization has the best track record of predicting who will win the presidency?
From: Tara C. of Grand Rapids, MI
Date: October 24, 2004 [revised November 2, 2004]

Gleaves answers: Eat crow, Gallup. Move over, Zogby International. Eat dirt and die, NBC/Newsweek. You don't even come close to being as good as WRC readers when it comes to predicting who wins presidential races.

As good as who?

The Weekly Reader Corporation (WRC) publishes a newspaper for school kids, and in every presidential election since Dwight Eisenhower's re-election it has invited our youngest citizens to predict who will win the November contest. Since 1956, the WRC poll has correctly dubbed the winner in 11 of 12 contests.

That's saying something, considering some of the close presidential elections in the last half century. In 1960 school kids correctly predicted that Kennedy would come out on top in a breathtakingly close contest with Nixon. Same with the fiercely fought battle in 1976 when incumbent President Gerald Ford was eventually overcome by Jimmy Carter, and the bitter contest in 2000 between George W. Bush and Al Gore that held the nation in suspense for more than a month.

The only election the kids got wrong was 1968, when they thought Robert F. Kennedy would beat out the Republican nominee. However, that survey was gathered in the spring of '68, months before the election and before RFK was assassinated in June.

Whom do school kids think will win in 2004?

The news release available this morning from Weekly Reader opens: "The students who read Weekly Reader’s magazines have made their preference for President known: they want to send President Bush back to the White House.... Hundreds of thousands of students participated, giving the Republican President more than 60% of the votes cast and making him a decisive choice over Democratic Senator John Kerry."[1]

It was almost an electoral sweep at every level. Elementary school kids in every grade voted overwhelming for George W. Bush. Among middle school kids the president also won, but by a narrower margin. Most high schoolers also preferred President Bush; only 10th graders voted in greater numbers for Senator Kerry.[2]

Besides the Weekly Reader poll, other indicators have traditionally presaged who wins in November.

For instance, The stock market's performance in the two months leading up to an election can tell you who will win. There have been 26 elections since 1900. In 16 of those elections, the Dow Jones industrial average trended up in September and October, and in all but one of those 16 elections, the incumbent party candidate won in November. In 10 elections since 1900, the Dow trended down in September and October, and in all but one of those elections, the incumbent party candidate lost in November. What is more, no president running for re-election has ever lost if the Dow in October is up at least 3 percent compared to one year earlier. But no president has been re-elected if the Dow in October is down by 5 percent of more, according to Jeff Hirsch in the Stock Trader's Almanac. [The less than stellar performance of the Dow Jones Industrial Average in September and most of October would seem to favor Kerry; however, there was a surge of earnings at the end of October, which might have reflected confidence in a Bush victory.]

Moreover, for the last 40 years the road to the White House has gone through the sunbelt; every winner since 1964 has been from the west or the south. Going further back, to 1948, candidates who came from sunnier, warmer states -- a home base to the south or west of their opponent's home base -- tended to win the White House. So:
- 1948: Missouri (Truman) beat New York (Dewey).
- 1952: Kansas (Eisenhower, who was actually born in Texas) beat Illinois (Stevenson).
- 1956: ditto
- 1960 is the clear exception to the rule: Massachusetts (Kennedy) beat Southern California (Nixon).
- 1964 saw two sunbelt contestants, as Texas (Johnson) beat Arizona (Goldwater); in this case, the candidate from the state with both western and southern elements won.
- 1968: Southern California (Nixon) beat Minnesota (Humphrey).
- 1972: Southern California (Nixon) beat South Dakota (McGovern).
- 1976: Georgia (Carter) beat Michigan (Ford).
- 1980 saw two sunbelt contestants, as Southern California (Reagan) beat Georgia (Carter).
- 1984: Southern California (Reagan) beat Minnesota (Mondale).
- 1988: Texas (with more than a touch of New England in George H. W. Bush) beat Massachusetts (Dukakis)
- 1992: Arkansas (the unambiguously southern Clinton) beat Texas (the ambiguously southern Bush who, remember, had New England roots).
- 1996: Arkansas (Clinton) beat Kansas (Dole).
- 2000: Texas (Bush) beat Tennessee (Gore).
- 2004: [The trend favors Bush of Texas over Kerry of Massachusetts.]

But don't count northern states out for their usefulness in determining the winner. Watch, for example, how the state of Ohio leans. Republicans have never won the White House without carrying the Buckeye State. [Bush is leading slightly in Ohio.]

Also, look at the "right track" or presidential approval poll numbers for the incumbent. If the last sizeable, reputable poll before the election shows that more than 50 percent of likely voters believe that the nation is on the right track, or that the president is doing a good job, then that is a common-sense sign that the incumbent will win. [Bush is at or slightly above 50 percent in most polls.]

And -- this one's really curious -- watch how the Redskins football team does in its last home game prior to the election. If the Redskins win, the incumbent's party stays in; if the Redskins lose, the incumbent's party loses too. This uncanny coincidence has prevailed for 17 straight elections -- all the way back to 1936. So:

1936 -- [Boston] Redskins beat the Chicago Cardinals 13-10; Democrat Franklin Roosevelt was re-elected.
1940 -- Washington Redskins beat the Pittsburgh Pirates 37-10; Roosevelt was re-elected.
1944 -- Redskins beat the Cleveland Rams 14-10; Roosevelt was re-elected.
1948 -- Redskins beat the Boston Yanks 59-21; Democrat Harry S. Truman was elected.
1952 -- Redskins lost to the Pittsburgh 24-23; Republican Dwight Eisenhower was elected.
1956 -- Redskins beat the Cleveland Browns 20-9; Eisenhower was re-elected.
1960 -- Redskins lost to the Cleveland Browns 31-10; Democrat John F. Kennedy was elected.
1964 -- Redskins beat the Chicago Bears 27-20; Democrat Lyndon B. Johnson was elected.
1968 -- Redskins lost to the New York Giants 13-10; Republican Richard M. Nixon was elected.
1972 -- Redskins beat the Dallas Cowboys 24-20; Nixon was re-elected.
1976 -- Redskins lost to the Dallas Cowboys 20-7; Democrat Jimmy Carter was elected.
1980 -- Redskins lost to the Minnesota Vikings 39-14; Republican Ronald Reagan was elected.
1984 -- Redskins beat the Atlanta Falcons 27-14; Reagan was re-elected.
1988 -- Redskins beat the New Orleans Saints 27-24; Republican George H. W. Bush was elected.
1992 -- Redskins lost to the New York Giants 24-7; Democrat Bill Clinton was elected.
1996 -- Redskins beat the Indianapolis Colts 31-16; Clinton was re-elected.
2000 -- Redskins lost to the Tennessee Titans 27-21; Republican George W. Bush was elected.
2004 -- Redskins lost to the Green Bay Packers 28-14.... [The pattern suggests Kerry will win, eh?]

There are other "signs" that are watched to predict the presidential race -- like the Iowa Electronic Futures; like Nickelodeon viewers' preference (there the kids accurately picked the winner from 1988-2000); like the top sales of Halloween masks of the candidates (sales of Bush masks are selling 10 percent better than Kerry masks this fall), to name just three. These offbeat "polls" are considered by many to be eerily accurate. But because of all the contradictory signals this year, all bets are off. What we know for certain is that some of the traditional "reliable predictors" are going to be wrong. Ultimately, the one poll that counts will be taken on November 2, when the ballots are counted.




[3]The pattern holds for the team specifically called the Redskins, whether in Boston (during the 1936 election) or in Washington (since the 1940 election). Interestingly, the Boston team had been called the Braves until 1933, when the name changed to the Redskins. Source: USA Today, November 1, 2004, p. 3C.

No comments: