Largest states are often best predictors in presidential elections

Posted Oct. 23, 2012

School of Communication
University of Miami

The presidential election takes place in a few weeks and the question on everybody’s mind seems to be, “Who’s going to win?”

It’s the question that has been all over every newscast for the past couple of months, the question that has been speculated on for countless hours by relentless talking heads, analysts and experts. Everybody seems to want to know before it happens.

With the unnerving amount of polls and information constantly being produced, people can feel that this answer is hard to come by and that looking into the future to predict the election seems impossible.

What if it wasn’t impossible to predict the elections? What if there was some simple way to tell who was going to win the presidential election the majority of the time? Wouldn’t that simplify things? What if finding out who was going to be the future leader of the free world could be achieved by looking into the past for answers, at voter demographics, state voting records, and exit poll results? Wouldn’t that seem to save everyone a lot of time and hassle?

Well luckily, there is a way to predict the upcoming election. And, all things considered, it can be mind-numbingly easy to figure out. There are multiple ways we can look at the upcoming election from a statistical standpoint to attempt to analyze it and these approaches will be further explored, but first let us look at the easiest, and most effective way to way to forecast the upcoming election: by looking at the voting trends of the largest population states.

America has an interesting election system. The Electoral College effectively decides the election, with states being assigned a number Electoral College delegates proportionately to the states’ population. Currently the amount of Electoral College votes needed to win the presidential election is 270.

Why is this relevant?

Because our Electoral College follows a winner-take-all system, so states like California (55 electoral votes, equivalent to slightly more than 20 percent of the votes needed to win) are heavily rewarded. Case in point: Since 1800, there have only been 10 elections in which the candidate who won the largest electoral state failed to win the entire election, according to statistics compiled from the U.S. Census Bureau website and the political website That represents a difference of only 17.5 percent.

In more simplistic terms, that means whoever wins California in this election (a recent University of California poll showed Obama with a 24 percent lead), from a historical standpoint, has an 82.5 percent chance of winning the entire election. If you had money on Obama winning in this year’s election, odds are from a historical standpoint, you made a good bet.

This approach though, is seemingly too simplistic. Surely predicting the winner of the upcoming election isn’t as easy as relying on who won the largest state, right? Perhaps it would be more intuitive to look at the demographic with the highest participation rates to see what influence it has on elections over the years.

Historically speaking, the demographic with the highest participation has been senior citizens. Since 1972, the 65+ age group has had the highest participation rates in every election, according to the U.S. Census. In the 10 elections since 1972, the senior citizen vote has correctly predicted the presidential winner eight times and incorrectly predicted the election outcome only twice. It doesn’t take a mathematician to realize that represents an 80 percent success rate in predicting the presidential election.

But, according to Christopher Mann, a political science professor at the University of Miami, those analyzing the election should stay away from trying to use only one demographic to predict an election.

“Look it just doesn’t make sense to try and pinpoint one state or one demographic as always forecasting the election,” said Mann. “Sure, more times than not whoever wins California or the senior vote takes home the entire election, but labeling one demographic as the be-all-end-all is futile.”

Instead, Mann argues that each election has its own key demographic and those demographics are constantly changing.

“One of the key demographics in this election as many have pointed out, and one that will continue to be pivotal in future elections, is the Hispanic vote,” he stated.

What’s most important in predicting elections, according to Mann, is to be able to study the progressive trends of certain demographics.

When we look at voting trends in recent years for certain demographics, we can see drastic changes. For instance, the all-important Hispanic vote has been moving away from the Republican Party, with only 31 percent of Hispanics voting for McCain in the 2008 election, down from 44 percent in the 2004 election, according to statistics compiled at the University of Connecticut’s Roper Center for Research.

Also, the power of the Hispanic vote has been simultaneously increasing.  According to the Pew Research Center, the number of eligible Hispanic voters has increased from 7.7 million in 1988 to 19.5 million in 2008 and is currently at 23.7 million for the current election. Not only that, but the percentage of Hispanic voters participating in presidential elections has gone up 4.9 percent since the 1996 election.

The growing number of both eligible Hispanic voters and the increase in the voter participation for Hispanic demographics is what makes the demographic so vital to this and future elections.

According to Joseph Uscinski, another political science professor at the University of Miami, these two factors are what decide the most important demographic.

“The two things you are looking for in key demographics is an increase in voter participation numbers and an increase in the total amount of eligible voters. The Hispanic vote currently represents both of those things.” Uscinski noted.

That is why, perhaps the best way, to predict the 2012 election, will be to follow the Hispanic vote. Hispanic voters have always had a low participation and registration rates in comparison to other groups, but have shown a continual increase in voting percentages and have shown a rapid increase in population, meaning there is more potential to gain votes from the Hispanic demographic than ever before.

According to William H. Frey, a senior fellow at the Brookings Institution for Public Policy in Washington D.C., the participation of minority voters is more important than ever.

Frey noted that not only have minority groups had lower voting and registration rates traditionally, but they also have had a smaller amount of eligible voters, noting that for every 100 Hispanic residents in the United States only 44 are eligible to vote. In contrast, he notes, 78 out of every 100 white residents are able to vote. Meaning that the latter will always be more heavily represented in elections. That’s why, Frey states, a predicted increase in eligible and participating voters is what will make the Hispanic vote so valuable.

“Minorities mattered in 2008 for three reasons: first, their relative sizes compared with whites increased in each state; second, their enthusiasm for the Democratic candidate was greater than in 2004; and third, white support for the Republican candidate (John McCain) waned in comparison to the previous election,” said Frey in a recent article for Brookings. “The 2012 election will most assuredly be a battle of turnout and its outcome will greatly depend on the enthusiasm of minority voting blocs.”

Based off of statistics compiled from the Pew Research Center and the U.S. Census Bureau, the Hispanic vote looks as though it will end up going Obama’s way, which might very well be the tipping point in this election.

Leave a Reply