June 10, 2016

Turnout was high in the 2016 primary season, but just short of 2008 record

More than 57.6 million people, or 28.5% of estimated eligible voters, voted in the Republican and Democratic presidential primaries that all but wrapped up Tuesday – close to but not quite at the record participation level set in 2008.

After a long decline, primary turnout reboundsFor a while it looked like this year’s primaries, driven by high turnout on the Republican side, might eclipse the turnout record set in 2008, when 30.4% of voting-age citizens cast ballots. The GOP did indeed have the highest primary turnout since at least 1980, according to our analysis – 14.8%, compared with 11% in 2008 and 9.8% in 2012. But turnout fell off markedly after Donald Trump won the May 3 Indiana primary and his two main rivals dropped out of the race.

Turnout in the first 29 GOP primaries – up to and including Indiana – averaged 16.6%, according to our analysis. But turnout in the final nine contests, after Trump became the presumptive Republican nominee, averaged only 8.4%.

By contrast, the heated battle between Hillary Clinton and Bernie Sanders engaged Democratic voters for longer. Less than a day before the final five states held their primaries, the Associated Press reported that Clinton already had clinched the nomination. But turnout in the five primaries held June 7 didn’t appear to be affected, averaging 14.1% compared with an average of 14.5% for the preceding 31 contests. (The District of Columbia will hold its Democratic primary June 14, but there likely won’t be enough voters there to significantly change the nationwide numbers.)

The overall Democratic turnout of 14.4% was well below the record 19.5% in 2008, but it was still the second-highest since 1988’s primary season.

All told, 39 states held primaries for one or both major parties this year (38 Republican and 37 Democratic, including next week’s D.C. primary), with the remaining states using caucuses or conventions to select national party delegates. For each primary state, we calculated turnout by dividing the number of votes reported cast in the party primaries by the estimated number of voting-age citizens (derived from our analysis of Current Population Survey data). We began with 1980, because primaries didn’t become a significant part of the nominating process until after 1968 and available turnout data for 1972 and population data for 1976 were incomplete.

A few caveats: We didn’t try to measure turnout in caucus states, because caucus attendance isn’t always reliably recorded and reported. Puerto Rico did hold Republican and Democratic primaries (the other U.S. territories held caucuses), but we didn’t calculate turnout because the CPS doesn’t include Puerto Rico.

Also, because states don’t always hold both Democratic and Republican primaries, the turnout rates for the two parties individually may not equal total turnout. In most states, total turnout was calculated as the total number of votes cast in the Republican and Democratic primaries divided by the estimated voting-eligible population; in states that held only one party primary, the vote in that primary was used to calculate turnout.

Topics: U.S. Political Parties, U.S. Political Figures, 2016 Election

  1. Photo of Drew DeSilver

    is a senior writer at Pew Research Center.


  1. Mari Hyatt1 year ago

    No mention of the rampant election fraud in the Democratic primaries, which obviously would make the 2016 #’s pretty much…inaccurate.

  2. Anonymous1 year ago

    I was hoping the voter turnout would be higher as the stakes are higher due to the presumptive candidates. Perhaps apathy has is setting in once again because of the candidates who remain. No matter what 28.5% of eligible voters is not enough.

  3. Daniel Se Kwon1 year ago

    Using the primary data, i ran couple scenarios for the upcoming election.
    Total Republican votes 29M. Trump only 13.5M
    Total Democrat votes 27.8M. Clinton only 15.8M
    If I run multiple scenarios, below results for electoral vote.
    Scenario A – No shift in parties (most likely not gonna happen)
    Trump 300 Clinton 238 (Trump President)
    Scenario B – 10% shift on both parties.(for example, Reps voting for clinton or Dems voting for Trum.
    Trump 285 Clinton 253 (Trump President)
    Scenario C – 20% shift on both parties
    Trump 290 Clinton 253 (Trump President)
    Scenario D – 30% shift on both parties
    Trump 274 Clinton 264 (Trump President)
    Scenario E – 10% shift only in Rep
    Trump 244 Clinton 294 (clinton President).
    Looks like a very tight race this upcoming November! I know Primary results means nothing but fun statistics.

  4. Anonymous1 year ago

    I strongly disagree with your statement in both this article and you earlier March 8th piece that primaries were not a significant part of the process until after 1968. My family was deeply involved in three Democratic Presidential primaries: Ohio 1952, JFK campaign 1960, and RFKennedy campaign in 1968. Furthermore, my late husband, Fred Dutton, was a member of the 1968-1972 DNC McGovern “Reform” Commission. We both worked for the JFK Presidential campaign, and Fred was Bob Kennedy’s de facto campaign manager. MY father, later US federal judge Timothy S. Hogan was a senior campaign advisor and chair of the Ohio primary for Kefauver I. 1952.

    Certainly there fewer primaries than today’s 38 state primaries and the District of Columbia (and the territories). But I would argue that inSpite of the numbers those primaries were ‘significant.’ History shows that Kefauver-Stevenson battles in 1952 determined the outcome. Would anyone want to make the case that if John Kennedy had not entered and won Wisconsin and West Virginia, he would have been the nominee? And even though the primary season was well underway before Bob Kennedy ever entered the race, his winning every primary after that, except Oregon, before his untimely death, made him the likely Democratic nominee. The people’s votes in the primaries were clearly significant– if not determinative–compared with the conventions and caucuses. Primaries even then generated enormous interest, too up significant financial and organizational resources of the candidate, not to mention time of the candidate, and generated enormous press. Far more press than caucuses and conventions and back room deals. The number of delegates selected by primary voters was smaller than the number chosen by ‘the party.’ True. But at least for the Democrats those and other primaries held before 1972 were historically very significant.

  5. Anonymous1 year ago

    It might also be helpful to compare a party’s participation rate to when they had sitting a sitting president running for reelection. In that light, participation by Democrats in 2016 is not remarkable when review against 1984, 1988, 1992, and 2008 (with 2000 and 2004 appearing to be more of the anomalies).