2008's 'What ifs...'

Plot your own demographic scenarios for the Electoral College results in 2008 — how race, gender and religion could change the outcome

By Jared Novack & Tito Bottitta / July 17, 2008
Small changes in a group's turnout and party preference can cause huge changes in the Electoral College results. To gauge these effects for yourself, based on what you expect at the polls on Nov. 4, adjust the target on the different grids to control how each group votes and which candidate can hit 270 electoral votes. Moving the target up-and-down changes the turnout level. Moving it left-to-right shifts the voting preference to either Democrat or Republican.
The map mixes Census data and 2004 exit poll results to calculate how shifts in a group's behavior could turn a state from red to blue (or blue to red). This is not an ironclad data analysis application, but a tool meant to examine the effect these groups have on the electoral map.


The map allows users to manipulate turnout and voting patterns to predict the results of the 2008 election. Using 2004 exit polls, the map establishes a baseline for demographic behavior. Combining this information with data from the Census Bureau and other sources, users are also able to predict what would happen if turnout increased or decreased in these groups.

The Census doesn't collect data for religion. However, using a combination of data from Pew (which recorded voting patterns from 2004) and the CUNY American Religious Identification Survey we can establish a similar snapshot.

2006 levels are based upon exit polls for House of Representatives voting.

There is one significant limitation to the map. Each voter falls within multiple demographic groups (gender, race, religion), but we don't have enough data to link the results. This allows the user to create paradoxical results; like what would happen if 100% of blacks voted for Democrats, but 100% of men voted Republican (what does that mean for black men?).

This is not an ironclad data analysis application, but a tool meant to examine the effect these groups have on the electoral map.

The small print

Turnout is based on the members of a group's citizen voting age population (which doesn't account for people who are ineligible to vote). Race and gender turnout is based upon the Census Bureau's November 2004 Current Population Survey. Racial groups are determined by White non-Hispanic alone, Black alone, Asian alone and Hispanic (of any race) groups. These are matched to the 2004 Exit Poll conducted by Edison Media Research via CNN. These groups are matched to the exit poll's categories of White, Black, Asian and Latino.

In many cases exit poll data is not available for a state because the demographic group is too small (eg. Latinos in Missouri). In these cases we've assumed the national numbers. Though this isn't an ideal situation, the size of unsampled groups is so small that the substitute data doesn't corrupt the overall picture.

Protestant denominations recorded by the CUNY survey are combined into a single "Protestant" grouping (excluding Mormon/LDS). Mormon data is based upon "Other Christians" designation from the 2004 Pew survey.

Some sample scenarios:

The white vote is still what counts

Though diverse, WHITES still dominate America's racial makeup. And because WHITES turnout disproportionately higher than other racial groups, they compose nearly 80% of the electorate. Adjust the turnout of other racial groups to absurdly high levels. Hypothetically, Obama's candidacy could motivate minority racial groups to turnout and vote Democratic like never before. It looks like he can't lose, right? Shift the WHITE voting preference slightly towards McCain. Now it looks like a Republican landslide.

The Asian vote?

Living in a city like Boston, it seems ASIANS are a significantly large racial group. But between the coasts, they represent a very small amount of the total population. In 2004 even if every Asian-American turned out to vote Democratic or Republican the election's result would not have changed.

Republicans are from Mars...

In 2004 men chose Bush with a 10-point gap, while women chose Kerry by a smaller margin. Because women are a larger part of the electorate, Democrats would need to motivate only a small slice to flip the election.

The forgotten religious group

You hear about the Evangelical vote, the Jewish vote, the Catholic vote even the Muslim vote (even though they represent only about 1% of the population). What about the Atheist/Agnostic vote? At about 12%, "None of the above" represents a significant chunk of the electorate. This group votes reliably Democratic, but turnout hovers around the 50% range. If turnout increases to the same level as Jews (high, but not unrealistic), it would have be enough to swing the election towards Obama's favor.