What We Think vs. What We Know


In attempting to forecast the upcoming elections, it is important to keep things in perspective, even if many others have lost their own. In relation to 2010 voting, wild predictions are still being made as to just how many seats will change hands in November. Unfortunately, many of these estimates may be based upon what people think, as opposed to what they actually know.

The late September primary results have complicated matters for Republicans as far as their chances for a takeover of the U.S. Senate. The Delaware result that led to At-Large Rep. Mike Castle becoming the latest victim of the Tea Party diminishes the Republicans’ chances of defeating the Democratic nominee in November, and the First State was a critical cog in any GOP Senate majority scenario. However, with the West Virginia and Connecticut Senate races become more competitive by the day, Republicans still stand a fair chance at winning back the Senate.  Additionally, many continue to forecast a Republican resurgence in the House big enough to exceed the 39 seats necessary to control the body in the next Congress. Some strategists site huge numbers. For example, former House Speaker Newt Gingrich said the number of Republican gains could exceed 75 seats. Such talk is not based in electoral reality, however.

Polling is the combination of statistical compilation and the views of a small number of people who are mathematically representative of the larger electorate. Thus, a poll
commonly cited as the foundation for many a political predictionis an example of what we think. There is no doubt that polling is a fundamental part of the forecasting process, but often times the data can lead to incorrect conclusions.  Polling is most valuable in forecasting trends leading to Election Day, not taking the results of one isolated survey and attempting to extrapolate a future result.

Utilizing tangible evidence is the more important part of projecting successfully. Looking at a district’s voter history and its primary turnout ratio are two examples of considering things that we know. Emphasizing this information leads to better conclusions because facts comprise our base hypothesis and not supposition. Naturally polls must be considered in attempting to predict the outcome of elections, but they should not be the exclusive
or even determiningfactor in understanding how the people of a certain district or state will eventually vote.

Here are some tangible examples of things we know:

1) In 1994, the last Republican landslide year, the Democrats were risking 28 open seats
obviously the easiest type of seat for the out party to convert. This year, Democrats put forth only 20 non-incumbent seats, sixteen of which are competitive. This means Republicans will have to defeat a minimum of 23 to 25 opposition party incumbents even if they win every competitive open Democratic seat, which isn’t likely. Republicans haven’t been very good at defeating Democratic incumbents. In the seven elections since 1994, Republicans have taken out a cumulative total of just 21 House Democratic incumbents, or an average of three per election cycle.  To automatically assume that the party can jump from three to 23 in one election may be a stretch. Where is the statistical basis for the conclusion to automatically assume that this can be accomplished when electoral history suggests otherwise?

The 2010 primary turnouts in a number of states greatly favor Republicans. In Indiana, South Carolina, Michigan, Missouri, Arizona, Florida, Tennessee, and Wisconsin, the number of people voting in the Republican primary far exceeded the Democratic total. This could be a harbinger of what is to come in November, particularly in states where such a trend is unusual, like Michigan, Missouri, and Wisconsin. If a Republican wave is to occur, look to these states to begin the trend.

Democrats, in their individual campaigns and within the national political party structure, are in the superior financial position. Outside group funding, however, appears to be favoring Republicans, thus neutralizing to some extent a key majority party advantage.

As part of the Republican majority strategy, seats like AR-1 and AL-5 must come through with GOP victories. Though the polls look good for these campaigns right now, a Republican has never won either of these districts. What tangible evidence is there to suggest that such a trend will buck history and finally change this year?

Another important House election predicting factor during this decade is the aggregate vote tabulation. The aggregate vote is the grand total for all contested seats; that is all districts that have a Democrat and Republican running against each other. The cumulative vote total represents a national Republican and Democratic percentage. In the decade of the 2000s, the party winning the aggregate vote has also won the House majority in every election since the 2001 redistricting process ended. This means the current maps are fair to the degree that they accurately depict the national mood. In 2008, Democrats received 54.1% of the aggregate vote, leading to a gain of more than 20+ seats, on top of the 30 they won in 2006. Republicans would be in range of majority control if they can exceed 52% on the 2010 aggregate vote scale.

These are just five factors that can give us tangible clues as to what we can expect in November. There are many more. Thus, in listening to pundits’ predictions about the fall elections, pay more attention to the ones who base their observations upon what they know as opposed to what they think.

Spread the Word

Help educate your friends on the importance of being civically involved.