by Michael A. Smith, Professor of Political Science, Emporia State University
It’s election time! This means plenty of excitement and trepidation. Some of it is valid, according to Political Science. Much of it is not. How can we use the tools of our discipline to separate the analysis from the hype?
#1 Enough Already with the Horse Race. The Polls are Tied.
That’s right, the election is a tie, both nationally, and in the key states whose electoral votes will probably decide our next President. This is because the polls are within the margin of error. When tied within the margin of error, polls cannot predict a winner. Period.
Polls operate via sampling, not unlike the way water quality is tested. It would be absurd to remove all water in Lake Michigan and take it to the lab. Instead, samples are drawn, those are tested, and then inferences are drawn in order to estimate the quality of water in the lake from the sample results. Polling is similar, sampling public opinion instead of water quality.
The next step is drawing the inference from the polling sample to the whole population. This means that there is a margin of error, along with a confidence interval. The polling results are within an interval around the actual number. If the difference in support for the candidates is within those intervals, no winner can be predicted.
This does not stop the news media from trying to do exactly that. For months now, the media have reported small changes from poll to poll, still within the margin of error. Granted, pollsters today have a few more tools. The most notable is to amalgamate results of multiple polls using a formula. This technique was developed by Nate Silver, who repeatedly forecast that “Donald Trump has a path to the Presidency” in 2016, accurately defying many pundits. But, this requires a pretty deep dive. For the average election observer, the best thing to do is simply note that the polls have the candidates tied within the margin of error, which they means they cannot predict a winner. What they predict is that the election will be extremely close. That’s where we are, and there’s not much more to say. In 2024, the results have been stuck within the margin of error for months, and will probably stay that way through the election.
Polls can do far more useful things besides trying to predict the winner. For example, they tell us what is on people’s minds. As I recently editorialized, this election is interesting because many objective indicators show the country doing well during the Biden Administration. Setting aside the issue that Presidents cannot personally take credit or blame for everything that happens while they are in office, circumstances today would seem to predict that President Biden would be popular. He should’ve been in great shape for reelection. Yet Biden is unpopular, and many Americans seem to think the country is not doing well. It appears that many Americans still have a lot of anxieties about inflation, despite the fact that it has been under control for the past year and a half. Americans want prices to go back down to where they were in 2020, rather than having inflation being brought under control by leveling off the continued increase in prices, which the Federal Reserve already did. What voters want may well be beyond what any President and Congress can do. This forces the candidates to get quite creative with their proposals, pushing the boundaries of Presidential power. For example, in a free market economy, the President has little control over housing or grocery prices.
#2 Don’t Overestimate Political Campaigns
As Sasha Issenberg points out in The Victory Lab, political scientists in the late 20th century typically assumed the political campaigns had little to no impact on election results. This is because the results were driven by fundamentals like the state of the economy and whether or not the country is at war. If the country is at war, it matters if we are perceived as winning that war. The last of those two only come into play in certain elections, such as 1952 and 1968, when Harry Truman and LBJ declined to seek re-election because each believed he could not win due to the Korean and Vietnam wars, respectively. On the other hand, the economy is ever present. As Bill Clinton’s advisor James Carville so famously said, “it’s the economy, stupid.”
Issenberg notes some new thinking on this. For example, using these indicators, Al Gore should’ve won the 2000 presidential election. In fact, he did win the popular vote, but that win was fairly narrow, and he lost Florida’s electoral votes and the Presidency after the U.S. Supreme Court decision, Bush v Gore. This re-introduced the idea that political campaigns can make a difference.
That said, many political scientists still think the impact of campaigning is greatly exaggerated in Presidential, general elections. Campaigning may indeed matter– Issenberg focuses on the strategy known as microtargeting, which targets narrow demographics of voters. Still, fundamentals like the economy are usually the major drivers. In 2000, the Presidential election was ridiculously close, and campaigning may indeed have made the difference. Then again, the same could be said for the Electoral College, the Supreme Court, the butterfly ballot, Ralph Nader’s candidacy, Florida officials “caging” African-American voters, and just about anything else that could tip the balance in a national election decided by fewer than 600 votes in one state. Generally speaking, it’s best to start with fundamentals like inflation and unemployment. In fact, many of Issenberg’s examples pertain to primary, not general elections, such as Barack Obama’s defeat of Hillary Clinton in 2008. Primaries feature lower turnout and candidates not yet well-known to voters, so they are more prone to being influenced by campaign activity.
#3 Neither Conventions, nor Vice Presidential Nominees are Likely to Matter Much
Democrats are all abuzz about Harris’ running mate, Minnesota Governor Tim Walz, a remarkable figure that appears to have energized the Democratic Party, as has the Vice President herself. Excitement about Walz was apparent at the Democratic National Convention, and this probably helped with fundraising, though not nearly as much as Harris’ own nomination for President. It may also energize volunteers. But the bottom line is, voters are not going to vote for the Vice Presidential nominee. There is simply no election in modern times in which the Vice Presidential nominee ended up being decisive in the election results.
Consider 1988, when George H.W. Bush (“Bush 41”, or “Bush Senior”) faced off against Massachusetts Governor Michael Dukakis. Bush nominated U.S. Senator from Indiana Dan Quayle for the Vice Presidency. Like Bush himself, Quayle was prone to malapropisms, which often caused him to be ridiculed by Democrats, the news media, and entertainers. Dukakis chose Texas Senator Lloyd Bentsen, who famously criticized the young, boyish Quayle during their Vice Presidential debate, saying “You’re no Jack Kennedy”. It was common for people to speculate that if the Vice President was elected separately, Bentsen would easily rout Quayle. In the end, it was the Dukakis-Bentsen ticket that got routed, winning only 10 states and the District of Columbia. It should be noted that political scientist Richard Fenno studied Quayle while he was a senator. Fenno broke his usual silence on partisan politics to defend Quayle, arguing that Quayle had been an able, capable Senator. The ridicule was unfair. Nevertheless, the bottom line is that Bentsen’s popularity versus Quayle had no discernable impact on the election results.
Another example would be George W. Bush. Before becoming President after the closest Presidential election in U.S. history, “W” figured out that the Vice Presidential nominee is not going to determine the winner. He chose the acerbic Dick Cheney from Wyoming, a state with only three electoral votes, which always go Republican. Bush’s choice of Cheney said a lot about the kind of Presidential administration he would lead, but it had nothing to do with the election results.
There’s plenty to say about Tim Walz, not to mention Republican VP nominee J.D. Vance, whose book Hillbilly Elegy I recently read and found fascinating. However, one thing we cannot say, is that their nominations will have anything to do with the election results. If you want to know the latter, focus on how voters feel about inflation.
The same as true of political conventions. These are a lot of fun for the party faithful, news media, and fans of politics. It is true that polls seem to reflect a so-called “convention bounce” for the nominee of whichever convention just happened. However, the bounce always fades, and polling results go back to pretty much what they were before, which is again driven by the economy and sometimes a few other factors, as discussed above. Nor does it matter where the convention is held—a fact that obviously influenced the Democrats to hold this year’s convention in Chicago, Illinois: a safely “blue” state. Conventions can be very useful in energizing the party base to give campaign contributions and volunteers. They may also influence a few undecided voters, which could be important in an extremely close election. Still, it’s important not to make too big of a deal out of these conventions. They are primarily pep rallies. Walz took that metaphor literally this year, inviting his high school football team up on stage with him.
For fans of politics, elections can be a lot of fun. For political scientists, they are gold mines of data. For all of us, they are a sobering reminder of just how much is at stake, for example regarding the events of January 6, 2021. Be that as it may, the news media appear to have recycled some of their old frames for election analysis this year: the horse race, misreporting poll results, and getting all excited about things like political conventions and Vice Presidential nominees. It’s Political Science to the rescue: that sharp tool that cuts through nonsense, so we can see what’s really happening.
About the Author
Michael A. Smith is a Professor of Political Science at Emporia State University. He has authored or co-authored five books, the most recent of which is Reform and Reaction: The Arc of Modern Kansas Politics (co-edited with H. Edward Flentje, Kansas 2024). He has other academic publications as well, and also writes newspaper columns carried throughout Kansas as part of the Insight Kansas group and blogs for the MPSA. Michael appears occasionally on television and radio in Kansas and western Missouri to discuss state and national politics. He was an expert witness for the plantiff in the Bednasek v Kobach case, decided together with Fish v Kobach by the federal district court for Kansas in 2018. Michael teaches courses in American politics, state and local government, and political philosophy. He received his Ph.D. from the University of Missouri in 2000. Follow Michael on X (formerly known as Twitter).