Jewish World Review May 14, 2004 / 23 Iyar, 5764
Poll-gazing: Are the numbers moving?
The patterns of partisan support in the 2000 presidential election have been with us for an unusually long period of time. They are apparent as well in the 1998 and 2000 House elections, in which Republicans won the popular vote by a 49 to 48 percent margin. They are visible as well in the 2002 House elections, in which Republicans won by a slightly wider margin, 51 to 46 percent. The question is: are those partisan patterns still in place today?
To get a sense of the answer, I looked at the state polls of presidential preference as listed in the invaluable Real Clear Politics on May 9. I compared the average percentage margin for George W. Bush or John Kerry in polls conducted in March, April, and May with the percentage margin for Bush or Al Gore in the 2000 election. States where the difference was 2 percent or less I set aside: they seem to be following the 2000 pattern. What I found is that there were many fewer states where the difference was 3 percent or more in the 17 battleground states that Bush or Gore won by 5 percent or less than in the other states where one or the other won by more than 5 percent.
Let's look at the battleground states first:
These numbers suggest that Bush is ahead in Pennsylvania, which he lost last time, and behind in Ohio, which he carried last time. The results in the other states would be the same. This would be a net gain for Bush of 1 electoral vote.
What reasons could there be for Pennsylvania and Ohio to move in opposite directions? Ohio has suffered from job losses in manufacturing more than Pennsylvania. And Pennsylvania has a bigger mass of affluent suburbs, in the ring 50 miles around Philadelphia, in which Bush ran far behind previous successful Republicans. One hypothesis I have advanced is that Bush might run better among affluent suburbanites in 2004 now that his opponent is promising to raise taxes on high-income earners. Perhaps the Pennsylvania numbers are evidence of this.
Now let's look at the nonbattleground states:
There are no likely changes in electoral votes here. But those numbers do suggest some states may deserve battleground status. North Carolina has suffered from manufacturing job losses, particularly in the furniture industry, which are a lively topic of political discussion in the state. If Republicans fail to carry the furniture counties by the usual wide margins, North Carolina, which Bush carried by 13 percent, could be much closer. But Kansas, Oklahoma, and Utah seem out of reach for Kerry, as does Vermont for Bush.
Most of the states where the Bush margin in recent polls is larger than his margin in the 2000 election are states with major metropolitan areas. In such areas, as I pointed out in an article in U.S. News just before the 2000 election, Democrats in the 1990s made their biggest gains as compared with 1988. Affluent culturally liberal voters who voted for George H. W. Bush to keep their taxes down voted for Bill Clinton and Al Gore to affirm their cultural values. Do the gains in Bush margin in California, Maryland, Massachusetts, New Jersey, and New York (and in battleground Pennsylvania) indicate that such voters are going back to Bush?
Maybe, but I approach the question with great caution. Let's look at the results in each of these states to see what seems to be going on.
California. This state falls on this list solely because of the Survey USA May survey, which put Kerry ahead by only 46 to 45 percent. Other polls in March and April show Kerry 9 to 11 points ahead, very close to Al Gore's 11 point lead in 2000. The Survey USA result could simply be wrong, the 1 in 20 polls that polling theory tells us will produce a result more than 4 percent off the total population. I know I won't believe it until I see some corroborating results.
But the regional and ethnic breakdowns are worth a look. Survey USA has Kerry ahead 62 to 31 percent in the Bay Area/Northern Coast, as one might expect; in 2000 the area went 61 to 32 percent for Gore. But in the other three areas Bush runs ahead of 2000 and Kerry runs well behind Gore. Bush carries the Central Valley 55 to 34 percent (53-42 in 2000), carries the Inland Empire 53 to 35 percent (50-46 in 2000) and barely trails 45 to 48 percent in Greater L.A. (39-57 in 2000). In these three areas, Bush Survey USA looks more like Arnold Schwarzenegger 2003 than Bush 2000. The Survey USA poll shows Kerry with only small leads among Hispanics (49 to 38 percent) and Asian/Other (47 to 44 percent). These are far better showings for Bush than he made in 2000.
The Survey USA poll may be just wrong. But if I were running the Bush or Kerry campaigns, I would be checking my California numbers soon. Bush strategists have been weighing whether to advertise in California. If their poll numbers match Survey USA's, it might be a good idea. And if Kerry's numbers match Survey USA's, they might have to do what Al Gore wisely decided not do in 2000, and advertise in California also.
Maryland. This state elected a Republican governor in 2002, the first time since 1966. Bob Ehrlich won by winning unusually large margins in the Baltimore suburbs and the nonmetropolitan parts of the state; Democrat Kathleen Kennedy Townsend won by the usual large margins in the Washington suburbs and Baltimore City. Is there a similar potential for Bush? Maryland is an expensive state, since you have to buy Washington and Baltimore television (and most of the Washington buy is wasted on Democratic D.C. and Republican Virginia).
Massachusetts and New York. Kerry is going to win these states by very wide margins, but perhaps not as great as Al Gore did in 2000.
New Jersey. In early April, a Fairleigh Dickinson PublicMind poll showed Bush leading 48 to 44 percent. Nobody, including me, believed it. Al Gore carried New Jersey by 15 percent. An April 20 Rasmussen Reports survey showing Kerry leading 51 to 39 percent seemed to confirm that little has changed here. Now, an April 28-May 4 Star Ledger/Eagleton-Rutgers poll shows Kerry ahead 43 to 37 percent. This is a high undecided/others percentage, double those in the other two polls; if you eliminate half the undecideds as nonvoters, the result is 48 to 41 percent Kerry. Also, both the Fairleigh Dickinson and Star Ledger polls were of registered voters, a group usually less Republican than likely voters, as in the Rasmussen poll. I confess I don't know what's going on in New Jersey. Both candidates are already on the air in the Philadelphia market which covers about 30 percent of the state's voters. Will either campaign want to go on the air in the New York market, the most expensive in the nation?
So far, both candidates have limited their time buys to 18 or 19 battleground states, covering 34 to 36 percent of the nation's voters. If these polls are accurateagain, a big ifthey could bring into play states covered by the New York, Baltimore, Washington, Los Angeles, San Francisco, San Diego, and Central Valley media marketsanother quarter of the nation's population. That will present tough decisions for both sides. If these states do turn out to be in play, this election will look less like 2000 and somewhat more like 1988, when New Jersey, Pennsylvania, and California all voted Republican by narrow margins.
Every weekday JewishWorldReview.com publishes what many in Washington and in the media consider "must reading." Sign up for the daily JWR update. It's free. Just click here.