The Primary Puzzle


STATE LEGISLATURES MAGAZINE | October-november  2014

In this era of ideological extremes and intense partisanship, what role do primaries play in shaping election results?

By Louis Jacobson

Where primary elections are concerned, there’s more than one way to skin a cat. And for years, political thinkers have debated what effect the design of a state’s primary has on electoral results.

In this age of sharp partisan polarization—when primaries often determine who occupies the seat more than the general election does—the question of how primaries can shape results has become increasingly urgent.

High-profile congressional upsets in recent primaries—House Majority Leader Eric Cantor in Virginia and Senator Thad Cochran of Mississippi (although he later squeaked out a win  in the runoff)—have also drawn attention to the debate over which type of primary best reflects the will of the voters.

Some political reformers see opening up primaries as a way to curb the influence of the parties’ ideological extremes, which tend to dominate in closed primaries that are open only to registered party members.

But does wresting primaries from the control of only registered party members actually result in the election of candidates with more moderate views? Research suggests it’s, at best, an open question. Those who have studied the phenomenon say the hard evidence is under-whelming.

A Variety of Options

States have a range of possibilities when deciding on what kind of primary election to use. At one end of the spectrum is the closed primary, which permits only registered party members to vote.

On the other extreme is the open primary that allows all voters to cast a ballot in the primary of their choice, regardless of their own affiliation. Some states use this method across the board, while others use it only in certain circumstances. In some states, voters have to publicly declare which primary they’re voting in, while in others, they can make the decision in the privacy of the voting booth.

Then there is the “top-two” system, which generally puts all candidates on a common primary ballot, with only the top two advancing to the general election. Washington in 2008, and California in 2012, were the first states to institute this new format when voters approved ballot measures.

Louisiana uses a variation of the top two in the general election, where all candidates are listed on the same ballot. Those who receive 50 percent or more of the votes win immediately, but if no one gets 50 percent, the top two finishers face off in a runoff.

There are other hybrid formats as well. Semi-closed primaries occupy a middle ground by requiring party members to vote only in their party’s primary, but allowing independents to choose either.

For reformers, the appeal of open primaries is clear: They hold the possibility of diluting the power of each party’s most active (and often most ideologically hard-core) members by allowing other (and presumably more moderate) voters to have a say in who runs in the general election, when turnout tends to be highest.

Low Turnout

In 2012, the number of votes cast in U.S. House general elections was more than four times higher than the number of votes cast in House primaries. In U.S. Senate races, 3.7 times more votes were cast in general elections than in primaries. 

This year, turnout for primaries also appears weak: As of mid-summer, in the 25 states that had held primaries, voter turnout was 14.8 percent, down from 18.3 percent in 2010, according to a new report by the Center for the Study of the American Electorate.

The top-two format has the potential to increase primary turnout as well, since it ensures that voters in heavily Republican or Democratic districts will have at least two candidates from which to choose, rather than just one.  So far, the top-two system has not increased primary turnouts in California and Washington, but experts say it’s still too early to judge.

Efforts to Change

In the last two years, efforts were made in 19 states to change primary systems, either in the legislature or by ballot. Only two survived: Utah lawmakers amended the state’s unique convention primary system, and Oregon voters will decide if the state should adopt a top-two system on Nov. 4.

 Illinois was among eight states that saw failed attempts to switch to an open primary. Illinois Representative Scott Drury (D) says the main goal of his bill was “to increase voter participation.” The legislation stalled in committee, but Drury says he will continue to fight for it because “too often I hear that people don’t vote in primaries because they do not want to declare a party. That is a travesty and needs to be corrected.”

Attempts to switch to a closed primary took place in five states, and four states (including Oregon) saw efforts to adopt a top-two system. In three states, bills were introduced to allow split-ticket voting.

Utah lawmakers changed their system of nominating primary candidates at party conventions, which Republican activists have long dominated in the deeply conservative state. The aim was to elect more politically pragmatic candidates —”not just one person pushed into office by a select, small group of individuals,” Lane Beattie, head of the Salt Lake Chamber and a former Republican state Senate president, told the Los Angeles Times.

 The new law, effective in 2016, preserves conventions but gives candidates an alternative way to get on the ballot by collecting a required number of signatures. In addition, the measure allows independent voters to cast a ballot in any party primary. “The idea was to do something affirmative to increase citizen participation,” says Utah Senate President Pro Tem Curtis Bramble (R), NCSL president-elect.

Unclear Results

So how well do the more open primary systems do in encouraging moderation?

Some evidence suggests open primaries can make a difference, while other studies say they have no impact. The top-two system may have a somewhat stronger influence on electoral results, but experts say more data are needed.

In a 2003 paper, political scientists Karen M. Kaufmann, James G. Gimpel and Adam H. Hoffman studied state-level exit poll data from 1988 to 2000 and found that open primaries boosted participation by younger and more moderate voters.

And political observers from around the country offer some anecdotal evidence about how the shape of the primary system had an impact.

In Arkansas in 1990, for instance, then-Governor Bill Clinton was running for his second four-year term. He was unopposed in the Democratic primary, but the GOP primary attracted two candidates—Sheffield Nelson and Tommy Robinson. Robinson was perceived to be the greater threat to Clinton’s re-election, says Hal Bass, a political scientist at Ouachita Baptist University. Tens of thousands of Democrats voted in the GOP primary instead, likely providing the margin of victory for Nelson. Clinton was reelected.

 In Alabama, in 2010, the state’s open primary format allowed Democrats to vote in a hard-fought Republican runoff for governor. Democrats “played a big part in helping Robert Bentley defeat Bradley Byrne, since the teachers’ union was adamantly opposed to Byrne,” says University of Alabama political scientist William Stewart. Bentley went on to win the governorship.

Most studies, however, suggest that, if open primaries do have an impact on electoral results, it is small.

Jon C. Rogowski, a political scientist at Washington University in St. Louis, says his research has found that switching to a more open system  can generate more moderate candidates—but the effect is small.

Michael McDonald, a political scientist at the University of Florida, also found a limited effect in a study that uses a new database of candidate ideologies drawn from campaign finance data. The study found that moderate candidates do have an advantage in relatively open primaries—but only when they come from the state’s dominant party.

“This makes sense when you consider that a party’s members defect to vote in the other party’s primaries only when the other party’s candidates are the ones likely to win in the general election,” McDonald says.

And finally, other research suggests switching primary systems makes no difference on the type of candidate elected.

Political scientists Seth Masket of the University of Denver, Eric McGhee of the Public Policy Institute of California and three co-authors looked at all state legislatures over the past two decades to see whether there was a correlation between primary systems and legislators’ voting patterns.

“We really didn’t find any impact at all,” Masket says. “Legislators from open primary states are just as polarized as those from closed primary states. We even looked at states that switched systems to make their primaries more closed or more open. We still didn’t detect any real effects.”

The study suggests several possible reasons for the lack of impact, including the persistent strength of party activists and the possibility that there are simply fewer politically active independent voters than previously assumed.

Turnout Too Low To Matter?

Wellesley College political scientist Hahrie C. Han—whose research with David W. Brady of Stanford and Jeremy Pope of Brigham Young University has found similar results—suggested the lack of impact may stem from the fact that turnout in primary elections “is so low overall that the potentially moderating effect of an open primary is not there.”

Douglas Ahler of the University of California at Berkeley suggested another possible factor: “Voters simply didn’t know enough about the candidates to distinguish moderates from extremists within parties.”

Kristin Kanthak, a University of Pittsburgh political scientist who has studied the role of primary structures, agrees. “To take advantage of my ability to cross-over vote in an open primary, I have to know the names of the candidates and know which one I like,” she says. “Most of the time, voters just aren’t sophisticated enough to use these systems to make a big difference. Party remains an easy cue for voters, so voters are going to rely on that.”

University of Arizona political scientist Barbara Norrander says her research suggested that, if anything, the path to more moderate candidates may actually run through closed primaries, not open ones.

In states with closed primaries, she found, more moderate voters tended to register with a party, because not having a party affiliation entails a real cost—the right to vote in the primaries. By contrast, in open primary states, the parties tended to attract more hard-core activists and fewer moderates, since moderates could choose to be independent without losing their vote in the primaries.

A Modest Moderating Effect

The top-two system may have a somewhat stronger influence on electoral results. Anecdotally, experts suggested a few examples in recent years. In Louisiana, a little-known Republican, Vance McAllister, won a 2013 special election for a congressional seat after finishing second in the all-party first round. Had Louisiana used a closed primary system, McAllister likely would have been eliminated in the initial round, and the first-place Republican and the leading Democrat would have faced off in the general election.

In California, the most notable example occurred the first year the state used the top-two model, in 2012. In the 31st Congressional District, which covers San Bernardino County, Democrats held a 41 percent to 35 percent edge in voter registration, but four Democratic candidates split 49 percent of the primary vote, while two GOP candidates split the remaining 51 percent, says A.G. Block, associate director of the UC Center Sacramento.

As a result, the top two candidates for the November run-off were Republicans, even though the district, by the numbers, should have gone to a Democrat.

“In 2012, some of the most heated congressional races in the country occurred in the districts in California that pitted two candidates from the same party against each other,” says Rogowski of Washington University.

“In these cases, the ideological differences between the candidates tend to be quite small, so the election outcome hinges substantially on mobilization and the personal characteristics of the candidates. Whether that is ultimately a good thing is something policymakers and voters will have to decide.”

In California, some research has shown the top-two system has a modest moderating impact, though the data are preliminary and, in many cases, mixed.

Ahler, for instance, says research he conducted with his Berkeley colleagues Jack Citrin and Gabe Lenz found “little evidence that voters are choosing more moderate candidates when given the opportunity to do so on the top-two ballot.” Yet, at the same time, he added, research by Christian Grose of the University of Southern California has found that legislators’ voting records after 2012 have been less partisan.

Too Early To Know

Block advises caution when drawing conclusions about California. “Most political scientists I’ve spoken with say it’s far too early to judge the overall effect of top-two,” he says. “They take the view that it will take three or four election cycles before the public becomes accustomed to the system.” 

So, is moving toward a more open primary system a good idea for states?

The best argument, perhaps, is that open systems tend to be popular among voters. And with more citizens choosing not to identify and register with a specific party—and thus being shut out from important primary voting—the appeal of open primaries is likely only to increase. At least that’s what political observers from states as far flung as Georgia, Louisiana and Minnesota say.

“The open primary is quite popular in Texas, and there is no serious talk of reforms that would move toward either a closed or semi-open primary,” says Mark P. Jones, a Rice University political scientist.

But while most observers see no major harm in a state switching from a closed system to something more open, they warn that states should be realistic about how much such a change can accomplish.

Based on the evidence so far, Rogowski says, “Changing primary systems is not likely to dramatically reshape the ideological tenor of the state’s politics.”

Kanthak, from the University of Pittsburgh agrees. “States should not see primary system changes as a panacea.”

The Variations


Closed primaries or caucuses allow only registered members of the political party to participate in the nomination process. Proponents believe only those committed enough to the party to register should  be allowed to decide who will be their candidates. They say closed systems contribute to strong party organization. Opponents note that closed systems exclude independent, unaffiliated and third party voters from the important nomination process.


Open primaries permit any registered voter to participate, regardless of political affiliation, or lack there of. Proponents argue this system gives all voters the most choices. Opponents counter that it gives non party members the opportunity to inter with their party’s nomination process.


In this system, all candidates are listed on one ballot, but only the top two vote-getters, regardless of party, get to compete in the general election. Proponents say this system gives independent voters an equal voice and may help elect more moderate candidates from the two major parties. Opponents argue it reduces the ability of third-party candidates to get on the ballot and limits voters’ choices as well.


Many states use a primary system that falls somewhere in between open and closed. In a few states, voters choose which party to register with by choosing which primary to vote in. Some states allow unaffiliated voters to choose either party’s primary to vote in, while other states bar them from both. And in other states, the political parties decide whether to welcome voters who are unaffiliated or from another party to participate in their primary. 

Wendy Underhill, NCS

Louis Jacobson is deputy editor of PolitiFact and a contributor to Governing magazine.

Additional Resources