Independent and unaffiliated voters are not the same

Weighing arguments for electoral reform means taking the difference seriously

We hear all the time that closed primaries disenfranchise independent voters. This is based on a false equation of party identification with registration.

Party identification (“party ID” for short) taps people’s emotional attachments to political parties. These come from childhood socialization, impressions left by charismatic leaders (think of Ronald Reagan or Franklin Delano Roosevelt), and how long one’s social group has been tied to either major party.

Party registration (or affiliation) is strategic, not emotional. It reflects the candidates or parties they want to see on a general-election ballot. Almost all states let people choose candidates by voting in primaries. Some states use registration numbers to decide which parties get ballot status.

Now contrast the share of Americans who decline to identify with the share choosing not to affiliate. People routinely give different answers to these questions.

Party ID is measured on a 1-7 scale running from “Strong Democrat” to “Strong Republican.” Figure 1 below gives the nationwide distribution for 2022. Data come from the flagship Cooperative Election Study. Two-fifths of Americans identified “strongly” with either major party. Adding “weak” identifiers brings that to 63 percent. That leaves 37 percent identifying as independent. Some of these people “lean” toward either party. If we drop them, the independent population is 19 percent.

Party affiliation/registration is measured by asking on a survey, then verifying answers against official registration. Figure 2 gives the distribution from the same survey. Two percent report no party affiliation or decline to state one. Another three percent are registered with minor parties. These groups make up five percent of Americans—roughly a quarter of those claiming to be independent.

Does the ID/registration mismatch persist over time? Figure 3 uses the same survey to plot the correlation between the measures back to 2012. Leaners are included among independents, and minor-party registrants are grouped with the unaffiliated. If anything, the mismatch was even greater before 2022.

Now consider a hypothetical friend. They claim to be “independent” but register as a Democrat. When pressed, they “lean” toward the Democratic Party (a 3 of 7 on the party-ID scale). They usually vote for Democratic candidates but sometimes do so by picking Working Families instead. That’s an easy thing to do if you live in a state with fusion. The idea, they say, is to send a message without spoiling the Democratic Party’s candidate(s).

Many primary-reform advocates claim that flavors of “closure” prevent independents from participating. For example, last fall in Washington, DC, a campaign for ranked-choice voting (with separate party primaries that are open to unaffiliated voters) implored the electorate to “let independents vote” in primaries (emphasis mine). Yesterday afternoon, a major newspaper ran an editorial with the following two lines (emphases mine):

“And semi-open primaries would give independent voters, who have long been shut out of the primary process, a voice in the city’s most important races.”

And:

“Also controversial — but arguably even more important — is the move to allow independents to vote in the city’s primaries. Primary elections in D.C., which are publicly funded, typically decide who wins general elections, because the District’s electorate skews strongly toward Democrats. The more than 75,000 D.C. residents who are unregistered with a party are often denied any real say on who runs their city.”

Yet, as we have seen, many independents can vote in primaries because they already register with parties.

The movement for electoral reform reflects dissatisfaction with party politics as they are. I have no problem conceding this. Not long ago, I wrote a book whose mission was, in part, to warn experts of potential consequences of not taking this sentiment seriously. The warning was that it could attach itself to any number of ill-conceived reform ideas.

Nowadays, I am interested in why people decline to identify with parties.

Letting unaffiliated persons vote in party primaries strikes me as benign. I even can imagine how it might improve an electoral system that has fallen out of step with key segments of public opinion.

Many other reforms are imaginable, however, and my plea as a political scientist is to see them discussed with more precision. Getting a clear sense of “pros and cons” may depend on it someday.