Skip to Main Content

Voter Resources

Who's Winning? Who's Going to Win? Can Anyone Know?

Unfortunately, the short answer is "No"; while you will see many headlines about polls over the course of an election, which will claim to accurately depict which candidates are leading, no poll or prediction model can foresee exactly what will occur on Election Day. This is why it's important to actually vote instead of letting any poll or series of polls lure you into thinking an election's outcome is predetermined.

Polls can nonetheless be useful indicators of current public sentiment, as long as voters are aware of how to analyze and interpret the information presented. The Bloustein Center for Survey Research at Rutgers University recommends that all of the following things be considered:

  • Sponsorship of the survey [Who funded this poll?]
  • Dates of interviewing [What notable events took place around the time this poll was cast?]
  • Method of obtaining the interviews (in-person, telephone or [e-]mail)
  • Population that was sampled [Do the demographics imply that this audience was already skewed toward supporting a specific candidate?]
  • Size of the sample [Is it too small a number to generalize from, depending on how large the total population for this election is?]
  • Size and description of the sub-sample, if the survey report relies primarily on less than the total sample [How small or large is the margin of error?]
  • Complete wording of the questions upon which the release is based [Do any questions in the poll seem intended to provoke a specific response?]
  • The percentages upon which conclusions are based

Adapted from: Bloustein Center for Survey Research. "Assessing Poll Validity." Rutgers University. Accessed August 29, 2024. https://bcsr.rutgers.edu/assessing-poll-validity/.

For certain elections, you will also want to consider how broad or narrow the poll's coverage is intended to be. Is a nationwide poll in a presidential election something you can interpret useful results from, or will it be more valuable to consider poll results for individual "swing" states? Likewise, in a state's gubernatorial election, which might tell you more: a statewide poll, or polls for specific counties or cities with large population centers?