Testing the Electoral College: Question Wording Matters

Since the founding of the Republic, Americans have elected 44 men as president over the course of nearly five dozen presidential elections.  In just five of those elections, the winner of the popular vote was not also the winner of the Electoral College and thus was not elected president.  Most recently, Hillary Clinton won the popular vote over Donald Trump by nearly 2.8 million votes; nevertheless, Trump prevailed where it mattered – in the Electoral College. Both the mechanics and the reasoning behind this constitutional quirk are often misunderstood.  And, when the Electoral College is put to the test in a public opinion survey, the waters often get murkier.  To coin a phrase, the Electoral College has consequences, and so does the language used by survey researchers to gauge voters’ opinions toward it and the process by which we elect our presidents.

The relative rarity of a president being elected without winning the popular vote has spawned fresh survey research into the question, “How should we elect our presidents?”  Four different survey research organizations since the November election have put the question to adults or voters in different ways.  What emerges is a not-quite-clear picture of public opinion on this issue depending on how the question is asked.

Generally speaking, the shorter and/or simpler the question, the better for the Electoral College, or, at the very least, the numbers are closer:

  • Suffolk University/USA Today’s December 2016 survey of registered voters asked, “Do you believe the United States should change the Constitution so the president is elected by the popular vote, not through the Electoral College?”
    • Forty-two percent (42{09f965da52dc6ab4c1643a77bd40d1f729d807040cd8db540234bb981a782222}) opted for yes, change, while 50{09f965da52dc6ab4c1643a77bd40d1f729d807040cd8db540234bb981a782222} opted for no, don’t change.
  • Gallup’s late November 2016 survey recorded a huge jump in favor of the Electoral College since the question was asked following the 2012 election: “Would you prefer to amend the Constitution so that the candidate who receives the most total votes nationwide wins the election, or to keep the current system, in which the candidate who wins the most votes in the Electoral College wins the election?”
    • Adults (not voters in this case) came down almost evenly split, with 49{09f965da52dc6ab4c1643a77bd40d1f729d807040cd8db540234bb981a782222} opting to amend the Constitution and 47{09f965da52dc6ab4c1643a77bd40d1f729d807040cd8db540234bb981a782222} opting for keeping the current system. That is down from 62{09f965da52dc6ab4c1643a77bd40d1f729d807040cd8db540234bb981a782222} amend/35{09f965da52dc6ab4c1643a77bd40d1f729d807040cd8db540234bb981a782222} keep in just four years.
    • Immediately following the hotly-contested 2000 election, when Al Gore won the popular vote but failed to capture the necessary 270 electoral votes, 61{09f965da52dc6ab4c1643a77bd40d1f729d807040cd8db540234bb981a782222} of Americans favored amending the Constitution, while 35{09f965da52dc6ab4c1643a77bd40d1f729d807040cd8db540234bb981a782222} wanted to keep the current system.
    • Immediately following the 2004 election, the numbers were identical (61{09f965da52dc6ab4c1643a77bd40d1f729d807040cd8db540234bb981a782222}-35{09f965da52dc6ab4c1643a77bd40d1f729d807040cd8db540234bb981a782222}).
    • The shift in the 2016 data can be attributed to a significant change in Republican attitudes toward the Electoral College. In 2000, 46{09f965da52dc6ab4c1643a77bd40d1f729d807040cd8db540234bb981a782222} of Republicans/Republican leaners favored amending the Constitution, climbing to 49{09f965da52dc6ab4c1643a77bd40d1f729d807040cd8db540234bb981a782222} in 2004 and peaking at 54{09f965da52dc6ab4c1643a77bd40d1f729d807040cd8db540234bb981a782222} in 2012. In 2016, that number plummeted to just 19{09f965da52dc6ab4c1643a77bd40d1f729d807040cd8db540234bb981a782222}.


Conversely, the longer or more detailed the question language, the more respondents are inclined to support changing the system and electing presidents solely through the popular vote:

  • A McClatchy-Marist survey from December 2016 asked voters the following: “Thinking about future elections, do you think who wins the presidency should be determined by the Electoral College, where the electors or representatives from each state vote for the candidate who won the most votes in the state, OR, the popular vote, that is, the candidate who gets the most votes in the nation regardless of the state in which people live?”
    • Forty-five percent (45{09f965da52dc6ab4c1643a77bd40d1f729d807040cd8db540234bb981a782222}) opted for the Electoral College, while 52{09f965da52dc6ab4c1643a77bd40d1f729d807040cd8db540234bb981a782222} said popular vote.
  • Similarly, a CBS News poll from December 2016 asked the following of adults: “Presidents are elected by the Electoral College, in which each state gets as many votes as it has members of Congress and can cast all of them for whoever wins in that state. Do you think we should keep the Electoral College, or should we amend the constitution and elect as president whoever gets the most votes in the whole country?”
    • Forty-one percent (41{09f965da52dc6ab4c1643a77bd40d1f729d807040cd8db540234bb981a782222}) chose to keep the Electoral College, while 54{09f965da52dc6ab4c1643a77bd40d1f729d807040cd8db540234bb981a782222} said we should amend the Constitution.
    • This is a slight improvement from CBS’s post-2000 election survey, in which 39{09f965da52dc6ab4c1643a77bd40d1f729d807040cd8db540234bb981a782222} favored keeping the Electoral College and 57{09f965da52dc6ab4c1643a77bd40d1f729d807040cd8db540234bb981a782222} favored amending the Constitution.


Clearly, survey language matters.  The McClatchy-Marist & CBS polls are longer than their Gallup and Suffolk counterparts.  The number of words in each question bears this out (see above).  It’s possible that longer, more detailed explanations could leave respondents confused about which response category they preferred, resulting in the difference vis-à-vis their shorter counterparts.  However, the more detailed question language in those two surveys may in fact provide respondents with more information and paint a clearer picture for them.  Either way, what is clear is that the way in which the question is put to respondents matters in determining the outcome. And, barring more survey research into the question, we cannot know for certain where American voters/adults actually come down on this important issue, although we do know it’s not a “landslide” for either side.

Public Opinion Strategies