Welcome to FiveThirtyEight’s weekly politics chat. The transcript below has been lightly edited.
sarahf (Sarah Frostenson, politics editor): From voter roll purges in Georgia that overwhelmingly affect black voters to a North Dakota law that could make it harder for Native Americans to vote, voter ID laws have garnered a lot of media attention leading up to the midterms. But a study from Pew Research Center published this week found that most Americans (76 percent) still support requiring people to show a government-issued photo ID to vote, despite growing evidence of a partisan divide.
So, the question I have for you all today is threefold:
- What do we know about voter ID laws and how they shape elections?
- What evidence can we point to to show their effect on the electorate?
- How do we reconcile public opinion with voter ID laws?
danjhopkins (Dan Hopkins, political science professor at the University of Pennsylvania and FiveThirtyEight contributor): I’d say that voter ID laws come in different shapes and sizes, so we often need to break them down by asking: What kind of ID is required, for whom, and what happens when voters don’t present ID? I think those distinctions often get lost in the debates over these laws — but they can be crucial to their impacts.
julia_azari (Julia Azari, political science professor at Marquette University and FiveThirtyEight contributor): I’d add to that and say that some of these are ID laws, but some are registration laws. What we saw happening in Georgia with absentee ballots being thrown out because their signature didn’t exactly match their voter registration file is qualitatively different from even a strict photo ID law.
danjhopkins: Yes, Julia, definitely a key distinction. Some of what people have been talking about in Georgia has been about voter registration, and the Supreme Court decision earlier this year on Ohio also had to do with voter registration, not voter ID.
For example, if Ohio cancels someone’s voter registration because an individual hasn’t voted in a few federal elections and then fails to return a postcard, that’s a voter registration law. But if Virginia requires someone who shows up to vote without proper ID to cast a provisional ballot, then that’s a voter ID law.
julia_azari: So I guess I’d actually classify both North Dakota and Georgia as examples of voter registration law rather than voter ID law.
danjhopkins: (Except North Dakota is complicated because it technically doesn’t have voter registration …)
julia_azari: Right, so maybe we’re talking about voter eligibility.
In Wisconsin, we have same-day registration, but complicated voter ID laws.
So, in other words, localized elections are basically a goddamn mess.
danjhopkins: But on the topic of voter ID laws, several recent studies of different laws seem to converge on a very rough consensus: These laws disproportionately affect voters of color and older voters, but they tend not to have large enough effects to swing election outcomes in any but the very closest races. Of course, if someone is denied the right to vote, that’s important in itself, regardless of whether a different candidate would have been elected.
Recent political science research has also pointed to the very real possibility that these laws could produce short-term backlash effects — that is, they mobilize some of the same constituencies that are disproportionately affected.
nrakich (Nathaniel Rakich, elections analyst): Yeah, Dan, I find the idea of backlash pretty interesting. For example, a group called North Dakota Native Vote and Daily Kos combined to raise half a million dollars in the aftermath of the Supreme Court’s decision to let North Dakota’s ID law stand. That money is now going toward helping Native Americans get the proper paperwork to vote and transporting them to the polls.
sarahf: I guess I find it hard to believe. Voter ID laws that make it more difficult for certain populations to vote … actually encourages them to vote in greater numbers? Or at least it does at first?
danjhopkins: Well, Sarah, I think it’s partly that these laws often tend to operate on relatively small slices of the voting population. When several co-authors and I tracked down the provisional ballots cast in Virginia in 2014 due to lack of ID, the number was only around 500 statewide. Now granted, tens of thousands of people showed up to vote in Texas and Michigan without the requisite voter ID in 2016, which gets back to my earlier point about how voter ID laws can vary a lot. But again, the number of people affected rarely adds up to an order of magnitude that is large enough to swing an election.
julia_azari: The fact that these issues have become partisan is terrible for democracy in the long term, but possibly good for mobilization in the short term.
It’s not surprising that people are mobilized when they feel specifically targeted or threatened.
People can get demoralized or immobilized for a variety of reasons and not vote, but there’s all sorts of evidence showing that specific kinds of threats can actually serve as the catalyst for effective mobilization.
danjhopkins: Yes, Julia, definitely. Backlash effects from unpopular laws are likely to fade, but if the laws remain in full force, the deterrent effects will likely not fade.
And if we think about it strictly from a partisan viewpoint, the impacts of the laws may change over time as well. As voters age, they are less likely to have current driver’s licenses, so we may see the partisan impacts of these laws change as the electorate ages.
nrakich: And voting coalitions may change. Hispanic voters may get more Republican, for instance. So then voter ID laws could start to hurt the GOP as well.
julia_azari: One question I haven’t seen addressed in the literature — and, granted, the paper hasn’t been out that long — is whether the mobilization that occurs around voting rights is mobilization that otherwise would not have happened. Or maybe mobilization that might otherwise have happened, but around a different issue.
nrakich: Right, on top of motivating them with anger to be more likely to turn out, news coverage of a court decision or new voter ID law can also help educate people about what they need to bring to the polls. It’s not hard to imagine an equally small-but-significant (I’m thinking in the four digits) number of people voting because of that, counteracting the turnout-depressing aspect of the law.
danjhopkins: Not only that, but also state and local officials can actively try to educate the public, as Virginia officials did in 2014. They sent a mailer to thousands of voters who didn’t have a driver’s license, which may be one explanation for the muted impact of that strict voter ID policy.
nrakich: Chances are, though, if a Republican government really did pass a voter ID law with the intent of driving down turnout (particularly among more Democratic groups), they won’t take steps to actively educate voters.
Republicans have admitted on occasion that this is their intent. One Republican in Pennsylvania, for example, argued for voter ID laws in 2012 by saying that it was “gonna allow Governor [Mitt] Romney to win the state of Pennsylvania.”
danjhopkins: But the Pennsylvania law was thrown out by the courts.
I happen to know something about voting in Pennsylvania as a voter here, and you only need to show an ID the first time you vote in person at a precinct.
sarahf: I think I heard on the podcast that you were registered to vote in several different states, Dan. 😉
danjhopkins: Not several, but while most states have been quick to pull me off the rolls, one has decidedly not.
julia_azari: Elections scholar Rick Hasen wrote a piece pointing out that the country is going in different directions. Some states are making it easier to vote.
Others are … not doing that. And there’s a pretty evident partisan pattern.
danjhopkins: It’s worth pointing out that there’s solid political science research on which legislators and states tend to pass the restrictive laws. It’s typically Republicans, and it’s typically in competitive states with sizable black populations.
nrakich: But these laws were pretty uncommon and uncontroversial until recently. They really started to take off in 2011 … after Republicans had taken control of a bunch of state governments.
danjhopkins: It’s also important to note that the 2013 Shelby County v. Holder decision by the Supreme Court freed up a number of states and localities to modify their election laws and administration without getting approval from the Justice Department.
Indiana passed the first in the recent wave of voter ID laws in 2005 — and it was upheld by the Supreme Court in 2008. But after 2010, the GOP took control of a variety of statehouses, so it had majorities in places like Wisconsin to enact new policies.
julia_azari: I think the Supreme Court ruling is really what’s making the difference here.
The court ruled that it’s constitutional for the Justice Department to oversee voting laws but that the coverage formula stipulated in the Voting Rights Act was outdated and should be updated by Congress. But that hasn’t really happened.
I also have a hypothesis: Public opinion about voter ID laws will shift as more people are affected by this stuff.
nrakich: It’s already shifting. As I mentioned earlier, voter ID laws used to be bipartisan and not particularly controversial. But now that they have become primarily identified with Republicans, you see a partisan split opening up in public opinion (even if it’s not as big as those on issues like abortion).
Political scientists, there’s a name for this, right? When people follow elites for their opinion cues?
danjhopkins: Nathaniel, there’s definitely a name associated with that — it’s John Zaller. His 1992 book, “The Nature and Origins of Mass Opinion,” is all about that.
nrakich: Sounds like some good light reading. I have a pretty chill week coming up.
danjhopkins: Yes, everyone digs Chapter 4, “Coming to terms with response instability.”
julia_azari: I’ve considered getting a response instability tattoo but haven’t yet.
danjhopkins: Good thing this is a chat or I’d be showing off mine …
sarahf: Hahah, but that’s what I don’t quite get in all this — the public opinion factor. There’s pretty broad consensus among Americans that showing some type of voter ID should be required to vote, but there’s also evidence that Republicans and Democrats differ on what they think the underlying problem (or solution) is.

nrakich: To narrate the chart Sarah posted, it’s pretty striking (to me at least) that most of the fight over voting access policy is among politicians. Voters are generally in agreement: Large majorities want voter ID but also other policies that make it easier to vote, like same-day voter registration.
Ninety-one percent of Republicans support voter ID laws, but so do 63 percent of Democrats. Eighty-two percent of Democrats support felon re-enfranchisement — but so do 55 percent of Republicans.
danjhopkins: Yes, this is definitely a case where the public opinion surveys are on broader questions while the policy debates address very narrow ones. Like, what counts as an ID? Can I vote with a student ID from a public university? What about a handgun license?
julia_azari: It’s pretty sobering to realize that even our forms of ID are partisan.
sarahf: Aren’t there some specific initiatives on the ballot this November regarding voter ID laws in some states? What are they asking for there?
nrakich: Yep. Both Arkansas and North Carolina will vote on enshrining voter ID in their state constitutions.
On the liberal side, Maryland will vote on same-day registration, and Michigan will vote on a whole package of changes, including automatic voter registration and no-excuse absentee voting.
Oh, and Nevada is also voting on automatic voter registration.
danjhopkins: Has anyone seen any polling on those ballot measures? Given that the public is in general in favor of voter ID, I imagine they’d get substantial public support, although maybe (à la the earlier talk of Zaller) a lot of Democrats will vote against?
nrakich: A September poll in Arkansas found that its voter ID ballot initiative was leading 71 percent to 21 percent.
In North Carolina, the voter ID amendment leads 64 percent to 27 percent as of early October.
julia_azari: But it’s worth thinking about the ways in which public attitudes touch on paranoia about non-citizen voting — and, I would argue, are tied to a long history of attacking the voting rights of black Americans.
danjhopkins: Absolutely — we can’t have this conversation without acknowledging that it’s deeply connected with a long history of systematic efforts to disenfranchise black voters and other people of color.
Proponents of these laws often say that they’re designed to prevent in-person voter fraud, but that is super, super rare. One study found that voter impersonation is reported at about the same frequency as alien abduction.
sarahf: Voter fraud is increasingly framed as a matter of election security, though. Right?
julia_azari: Exactly, which is unfortunate, as there are some pretty significant problems with election security.
nrakich: Ironically, one of the more plausible ways that election security could be compromised is that hackers access voter files and delete voters from them.
julia_azari: Right. Which is also a pretty big problem with American democracy even without Russian hackers or whomever. Rather than having a problem where too many people are itching to vote, we have the opposite problem. Millions are eligible to vote but do not.
danjhopkins: Earlier Julia mentioned just how decentralized voting in this country is, and she’s 100 percent right — voting is administered by our neighbors willing to serve as poll workers, in very different ways in different places. And that cuts both ways. On the one hand, it means that our data is decentralized and the chance that things go wrong in a precinct is high. But it also means that the level of coordination it would take to hack our elections without our knowledge is massively, perhaps unimaginably, high.
julia_azari: It also means that the implementation of laws is dependent on poll workers knowing the laws and being reasonable. Not to disparage folks working at the polls — that’s real civic commitment. But it can go wrong.
sarahf: OK, so the evidence that voter ID laws can change the outcome of an election is … small, but there is evidence that these laws do affect whether people vote both in the short term and the long term. What should we be looking for on election night and beyond?
danjhopkins: The most credible studies of voter ID’s effects on turnout involve looking at individual-level data, often merged across different sources. So I don’t think there’s anything we’ll be able to see on election night. But in the weeks and months afterward, once we’ve got voter file data, we might be better positioned to tease out some of the individual-level impacts — that is, which groups look like they were deterred from voting where. In some cases, we’ll also be able to review the document trails. And to reiterate, the short-run impacts of enacted laws tend to be on a scale that would only tip incredibly close elections (think the presidential election in Florida in 2000). But it’s hugely problematic that specific individuals are still unable to vote as a consequence — and those individuals are disproportionately people of color and older voters.
julia_azari: At the risk of turning in my political science card, I’ll be looking for stories that suggest something more systematic might be going on at the polls — laws being unevenly implemented or used to intimidate voters.