Lessons IT Departments Can Learn from 2016 Presidential Election Data

No matter your politics, you can’t deny the drastic difference between the 2016 pre-election polls and actual election results.

A vast majority of polls indicated a win for Democratic candidate Hillary Clinton. But Election Day results named Republican candidate Donald J. Trump the winner.  

The disparity in results has led to a lot of questions…

Where did the 2016 presidential election data go wrong? What lessons can businesses learn from election data errors? More importantly, what can an IT manager do to ensure that the misinformation of the 2016 presidential election doesn’t happen at his or her organization?


Why Were the Polls So Wrong?

Endless polls predicted the results of last week’s presidential election. Most were wrong. Why? A number of data errors led the pollsters to the wrong conclusions.  

For one, there’s the data set error. Polls only gather a limited quantity of data. So, they don’t have all of the information necessary to understand what’s truly going on. As a result, the polls indicate guesstimates rather than cold, hard facts. This is because polls measure likely voters. And “likely” doesn’t necessarily guarantee a vote on Election Day.

Another source of error is static data. Data is constantly changing, but polls only capture a moment in time. Election forecasters try to look at the data from a trend perspective, asking the question: how has the data changed over time? Theoretically, this should give an accurate look at the big picture. But in practice for the 2016 presidential election, the forecasts were so very wrong because most of the polls were wrong.

Shifting voting practices played a role in data errors as well. Take the early vote, for example. A record number of voters voted early this year. This should have made the numbers more accurate. But it actually meant that turnout on Election Day was lower than expected. The polls anticipated more votes on Election Day, and they simply weren't there. On top of that, absentee and and mail-in ballots continue to be counted days after the election. 

Or you can consider the data errors from the perspective that the polls weren’t entirely wrong—they just came to the wrong conclusion. Polls indicated that Clinton would win and she did win—the popular vote. But not the Electoral College. The polls that actually went wrong were the polls that indicated states like Wisconsin, Pennsylvania, and Michigan would vote for Clinton—but ultimately voted for Trump.

Data errors kept most pollsters and forecasters from accurately predicting the results of the presidential election. But you can keep your business from making the same mistakes.

Keep up with the changing world of business intelligence >


7 Lessons for Your IT Department

There are seven essential lessons about data that your organization should learn from the 2016 presidential election.

1. Never Underestimate Your Data

The polls underestimated the turnout of Trump supporters, and that’s where they went wrong. Trump supporters turned out in greater quantities than expected, and Clinton supporters didn’t.

Your lesson: don’t underestimate your data. When you consider your data, take all potential scenarios into account. Think about each region carefully and factor the information into your decision-making. Don’t underestimate or discount any subset of data.

2. Prioritize Data Accuracy

The polls didn’t have the most accurate information available—and this was likely due to the lack of complete information. As a result, polling accuracy can be easily compromised by factors like the increased use of cell phones and increased difficulty to reach the right voter groups.

Your lesson: Data means nothing if it isn’t accurate. Your organization should take steps to ensure that data is accurate before factoring it into your decision-making. Here’s how: Gather all of the information available and vet it for accuracy. Consider a business intelligence solution to maintain accurate data and inform decision-making.

3. But Don’t Ignore Bad Data

Bad data is inaccurate data or out-of-date data. Pollsters may have received inaccurate data from responders. For instance, responders may have been dishonest about which candidate they would vote for, or they may not have voted at all. Or the responders might have changed their minds between the date of the poll and Election Day.

Your lesson: Bad data shouldn’t be simply ignored. It should be corrected and vetted to turn it into accurate data. Your organization needs a way to find the bad data, so you can correct it. That’s where a business intelligence solution comes in.

4. Consider All Data Points

The polls were wrong about Wisconsin, and this could be because they forgot about one key data point. The 2016 presidential election was the first election in which Wisconsin voters needed to present a photo ID.

Your lesson: Don’t neglect a data point. Each data point is a crucial contribution to the big picture. Your organization needs each and every data point in order to make informed decisions. After all, you never know which data point could change everything.

5. Focus on Live Data

New polls surfaced on a daily basis leading up to the 2016 presidential election. Yet the polls still got it wrong. Though polls were constantly conducted, they weren’t always acting on the most up-to-the-minute information.

Your lesson: Make real-time data a priority. Data is constantly changing, and you need to make sure you have the most recent data before you act on it.

6. Carefully Examine Data Trends

Data trends must be consulted, but not weighted too heavily. States like Iowa, Wisconsin, Ohio, Pennsylvania, and Florida broke from recent presidential election trends and ultimately led to Trump’s Electoral College win.

Your lesson: examine data trends carefully before hedging your bets. Historical data trends might give your business indicators for success—but data trends can quickly change. Enlisting a business intelligence solution can help your business both take a trend into account and examine real-time data.

7. Remember Your Data Quirks

It can be said that the 2016 polls were, in fact, accurate and well within their margin of error when it came to the national popular vote. The thing is, the U.S. presidential election has a very important quirk that ought to have been accounted for in the polls: the Electoral College.

Your lesson: don’t forget about your business quirks. When gathering your data, make sure to factor in any quirks that are specific to your business. Once you have that information, you can reach an accurate conclusion.


Brush Up on Business Intelligence

Bad data leads to misinformed decisions. Your IT department undoubtedly has a lot on its plate. But it’s more important now than ever to obtain and maintain accurate data. And that’s where smart business intelligence strategy comes in.


Get Up to Speed

Discover “4 Things IT Managers Should Know About the Changing World of BI” and successfully navigate business intelligence for your organization.