Princeton Election Consortium

A first draft of electoral history. Since 2004

Jul 12: Biden 381 EV (D+6.3% from toss-up), Senate 53 D, 47 R (D+5.6%), House control D+8.0%
Moneyball states: President AK AR IA, Senate MT KS ME, Legislatures KS TX NC

The impact of voter ID rules: Perceptions vs. reality

October 1st, 2012, 9:00am by Sam Wang

Many of you have asked about the effect of voter ID rules. We’re joined once again by Ed Freeland, Director of the Princeton Survey Research Center. He is kind enough to offer some thoughts on a recent survey. Welcome back, Ed. -Sam Wang

In many states, recently enacted rules (ProPublica) will require voters to provide some form of identification when they show up to vote on Election Day. Based on estimates in a recent report from the Brennan Center at NYU, the rules could affect both poll accuracy and real outcomes in key states like Pennsylvania and Florida. It’s a hot-button issue: The push for these laws has come largely from Republican legislatures, who are likely to benefit from reduced turnout. Democrats point to evidence that in-person fraud is nearly nonexistent.

Now we have data on its perceived effect, thanks to a new survey.In swing state poll results released Wednesday by Quinnipiac University/New York Times/CBS News, respondents were asked: “Will changes in your state’s voting and registration rules make it harder or easier for you to vote this year or won’t the changes make a difference?”

The results reflect the partisan nature of the question. In Ohio, 12% of likely-voter respondents said it would be harder to vote. But 22% of Democratic likely voters said yes, compared with 1% of Republican likely voters – an enormous difference. A similar story emerged in Florida (23% of Democratic likely voters said yes, and 3% of Republican likely voters – an 8-fold difference) and in Wisconsin (19% of Democratic vs. 2% of Republican likely voters – a 9-fold difference).

These figures exaggerate the probable partisan breakdown of voters at risk. Voters without IDs are a mixed bag: low-income people, whites without college education, retirees. The net outcome, reduced turnout, will tend to favor Republicans. But not by a ratio of 8:1 to 22:1. The final effect has been estimated to be typically less than 1% of swing.

Also, voters who are not able to provide sufficient ID at the polls on Election Day still have an option: they can cast provisional ballots. How those ballots ultimately get validated and tallied depends on each state’s rules. It is unlikely that they will make a difference for the Presidential outcome – but Senate and House races, as well as other state questions, could be affected. One example is Indiana, where the Senate race is close. If after the initial tally on Election Day, the number of provisional ballots is smaller than the margin of victory for any candidate or initiative, then no one need worry about how those provisional ballots are marked. Ultimately, the validated provisional ballots will be added to the final official tally in the weeks after Election Day.

For the Democrats, a much worse scenario is the potential for the new ID rules to discourage voters from going to the polls at all. If nearly 1 in 5 Obama supporters perceives the new rules as a barrier to voting, how many will just stay home on Election Day? Maybe that’s a good question for the next swing state poll.

Our thanks to Mike Kagay and the staff of the Quinnipiac University/New York Times/CBS News poll for including this important question in their most recent voter surveys. No doubt we’ll see more on this issue in the coming weeks.

Author: Ed Freeland
Survey Research Center
Princeton University


Tags: 2012 Election · Politics

40 Comments so far ↓

  • badni

    “The net outcome, reduced turnout, will tend to favor Republicans. But not by a ratio of 8:1 to 22:1. The final effect has been estimated to be less than 1% of swing.”

    That is a complete non-sequitur. If the effect is that 1% of D voters are turned away and 0.01% of R voters are turned away, that’s a 100:1 ratio and about 1% of swing.

    And, I don’t know why 1% is considered not a lot. Over the course of hundreds of elections at all levels, you can tip a lot of races that way.

  • Matt McIrvin

    There’s been a lot of fear that this could tip the presidential election, and if it’s a <1% effect it's pretty clear that it won't in 2012, barring something else dramatic happening.

    But there *will* be some elections swung by vote-suppression efforts this cycle, and I hope that when it happens people recognize it and there's a big stink about it.

  • Matt McIrvin

    Incidentally, I actually encountered a speedbump on the way to voting *in Massachusetts* in the September state/congressional primary: we have a law here that puts voters on an “inactive” list if they’ve forgotten to turn in their yearly town census, and, having fallen in this trap, I had to get off it by presenting ID that established residency.

    I have heard that some Tea Party groups around Worcester sent poll-watchers who made a big deal of challenging people, ostensibly to make sure that this law was followed. But there were reports of them giving misleading information about what kind of ID was required (it doesn’t have to be a photo ID; an electric bill is good enough) and causing some other trouble.

  • Lindsey K

    Dr. Wang, If you have not yet addressed this, could you please discuss the general “enthusiasm gap” between Republicans and Democrats, even without the voter ID issue? Of concern to me is the gap I see at the Rand ALP site that shows a 4% gap in favor of Romney voters.

    • Matt McIrvin

      Republican-leaning constituencies have been more likely to vote for a long time; the Democratic base includes young people, poor people and racial minorities, all of whom vote at lower rates than the average voter.

      But the RAND survey, as far as I can tell, already has this difference baked into its election prediction, which has Obama well ahead.

      There were reports early in this campaign that the “enthusiasm gap” somehow exceeded long-term trends and that the electorate might resemble the 2010 midterm, when Democrats disproportionately stayed home and Republicans were so fired up that the Tea Party wave happened. I think the only people who believe that now are the “unskewed” dead-enders.

    • wheelers cat

      @Lindsy, Matt
      I think voter enthusiasm may correlate in some sense with social media effect in 2012. This election is a New Event in many aspects, one of which is social media usage and internet access.
      For example, RAND measures change in a fixed population of internet users.

  • Anbruch

    If we really are in a knife’s edge scenario with respect to the House, the suppression effort could have an effect on the close House races. I would imagine that the suppression would have its greatest effect in districts where it would not matter much (urban, strong D), but that targeting Dem-leaning voting precincts in close House districts could have larger effect. I imagine that Dems will anticipate this and plan to have large numbers of poll watchers in such voting precincts. I know there is a lot of training going on right now to send poll watchers into PA and OH. Whether they are paying sufficient attention to the particulars of House races is the question.

  • MAT

    I sit on our County Board of Elections. One thing that we drill into our poll workers over and over and over again is if someone shows up to vote, they get to vote.

    If there is a problem with the registration data, lack of an acceptable ID (for the states that require one) or there is a challenge that isn’t overruled by the precinct judge – the worst case scenario is that the voter submits a provisional ballot. The validity of these provisional ballots will be ruled on by the various election boards, subject to legal review if neccessary.

    My expectation is on election day, if there are an abnormal amount of provisional ballots cast due to voter ID laws (the 1% scenario will fit this description), that there will immediately be legal challenges filed on the behalf of those voters. Given the tenor of comments put out by judges in some states (Penn, for example) concerning the voter id laws, I would anticipate these challenges to receive a sympathetic ear as actual, rather than theoretical harm to the plaintiff can be demonstrated. So my worst case scenario is that we wind up with a lot of Al Frankin situations, where the actual results are not known for some time due to uncertainty around the validity of those provisional ballots.

    • MAT

      The best advice I could give to members of any political persuasion – if you are worried about what some other group may or may not do on election day – volunteer to be a poll observer with your local party. Knowledge of what is happening out at the polls is the best way to prevent mischief.

    • Olav Grinde

      MAT, glad to hear you’re on the County Board of Elections! I have a few questions.

      First, I understand that state elections are usually called before the provisional votes are counted. Do you know of any instance where the results have been changed due to the provisional vote count? Has this happened without a court order?

      Also, I wonder if you could cast some light on provisional votes. In your count and state, how many provisional votes were there in the last election, and what percentage of voters had to cast provision votes?

      Do you happen to know what percentage of provisional votes were accepted/rejected?

      To what extent might challenges from poll observers / interested parties (say, from the Tea Party) influence on whether or not a voter has to cast a provisional vote? Can a voter challenge the challenge, so to speak, on the spot?

    • Paul K2

      The real game this election, will be the successful strategy used in Ohio in 2004, and attempted in the Wisconsin re-call election. Right wingers are preparing to challenge voters, especially at the last minute, and confuse the poll workers. Their hope is to slow the process down, so that the line builds. In Ohio in 2004, there were 8-hour lines in some areas around Cleveland and in the college towns.

      This year, a similar game is afoot; the right wingers have begun challenging voters based on address corrections. The most infamous of the current tactics involve challenging voters at universities for not marking their dorm room on the voter registration. If only the dormitory building address itself is used, the “vote fraud activists” try to throw the voter off the roles due to an improper address.

      Hopefully the swing state poll workers are wise to this maneuver, and simply have all the challenged voters fill out a provisional ballot, and keep the line moving, instead of getting caught up in the confusion.

      In my current state of residence, Washington state, we vote by mail. This has the effect of increasing voter participation, yet also seems to benefit rural voters (often Republicans), by making the trip into town unnecessary to vote. I am a huge fan of mailed ballot voting.

      When I look at what’s going down in my home state of Pennsylvania, I almost cry. The move to reduce voter turnout is blatant and transparently obvious. I hope it doesn’t work.

  • Zack

    I just moved to Pennsylvania (for reasons entirely unrelated to the election). I have observed two on-the-ground effects, pulling in opposite directions:

    On one hand, converting my CA drivers’ license to a PA license (legally required even if I wasn’t going to vote) cost me two full working days dealing with red tape. It wouldn’t have been quite so difficult if I had just wanted the “state ID” that enables voting but not driving, but I can easily see the working poor, in particular, not being able to spend the time even for that. (However, once I’d jumped through all the hoops, I was automatically registered to vote by the DMV’s computers.)

    On the other hand, here in Pittsburgh I have been seeing a lot of advertising (on buses and billboards) and people at tables, with the message: “Those Bastards in Harrisburg are trying to Steal Your Right To Vote. Don’t let them!” I don’t know who’s paying for it, and I don’t know how effective it is, but if that message resonates with the population that were targeted for voter-suppression tactics, I can imagine the whole thing backfiring on the state legislature pretty dramatically.

    Personally, the experience converted me from Disgruntled Likely Voter to Pissed Off Certain Voter, and means that I’m actually going to bother researching state and local races.

    • DaveM

      I’ve lived in Pennsylvania most of my life, and as far as I’m concerned, this Voter ID legislation is the most egregious abuse of the concept of representative government I’ve seen at the state level. From the Magna Carta to the Voting Rights Act, the franchise has been expanded and protected for 800 years; now a mediocre cabal of political hacks is taking that most vital right of citizens in a free society possess and trashing it.

      Okay, now that I’ve got that off my chest, I think the damage from such legislation is probably more long- than short-term. A class of folks, mostly poor, elderly, and/or nonwhite–whose interests are arguably already not being served by their representatives–are now being made to jump through an absurd series of bureaucratic hoops just to avoid losing what little influence they have.

      Under these circumstances, it’s hard to imagine that levels of participation among these folks won’t decline. I expect that, absent widespread overturning or repeal of Voter ID legislation, over time we’ll see that participation in the electoral process by those most affected by such laws will drop off by a much more significant percentage than the actual instances of outright disqualification.

  • MAT

    @Olav – First, in my state, provisional ballots are processed (i.e. a ruling is made as to the validity of the ballot) prior to canvass, which is when the election is certified. The process is pretty similar elsewhere. So you won’t have ‘official’ results prior to the provisionals being handled. If there are no races that are within the margin that provisional ballots could override, then there is no need to take any action on them. Otherwise, they are processed *prior* to certified results.

    2nd – the most recent well known instance of a race being changed by provisional ballots was the Al Franken/Norm Coleman Sen race in 2008 (btw, there is a fascinating discussion of this recount and how the two sides treated provisional ballot challenges in the excellent ‘The Victory Lab’ by Sasha Issenberg – highly recommended.)

    We recently had a state representative primary race that was within 7 votes on election day. After processing the provisionals, the gap actually widened to 11 votes. Obviously, it could have gone in the other direction. This was in the normal scheme of things, no court order needed, and occurred in the week between election day and when our ‘canvass’ occurred (canvass being when we total all the votes up and fill out the official paperwork to the state certifying the results). A court might step in if a voter felt that their provisional ballot had been incorrectly ruled invalid – the mechanism for this is usually spelled out in the state election law.

    Why is a provisional typical ruled invalid? Overwhelmingly it’s claims that a person is registered and they aren’t. For example, in our state you can register to vote when getting/renewing your drivers license. Sometimes there is a delay for that registration to make it to our rolls – however, we can check with the DMV in order to validate the registration request. If it exists, we’ll count the provisional ballot, otherwise no. If we can find that the voter is indeed registered somewhere, we’ll count the vote. Sometimes it’s a case of a person voting at the wrong precinct, name spelled differently on registration, etc…. Also the case Matt McIrvin referred to above, registrations may be marked as inactive, or even improperly removed from the registration rolls (for example, a person really isn’t dead!). Although most of these are fixed on election day, it could be corrected when the provisional is processed.

    In the elections I’ve been involved with, we normally have a very small provisional rate – at or below 1%. The numbers are so small that I can’t give you any sort of statistically significant number of how many are ultimately rejected – one election, it was all of them. The last primary runoff, we accepted about 1/2 of the provisionals. It all has to do with the individual circumstance around each ballot. BTW, we make a determination about the validity of the ballot before seeing the ballot itself (they are kept in envelopes sealed by the voter with the needed information on the outside). This way, you make a ruling on the ballot without any information on how that individual voted.

    Because lack of voter id would be a specific category for a provisional, I would expect any sort of significant numbers of these to kick off a legal challenge – for example, if you were a losing candidate in a close race – you’d see that stack of provisionals with people without id’s as your potential salvation. They’ll get challenged legally, believe me.

    As far as your last question – everyone has a right to cast a provisional ballot if denied the opportunity to cast a normal one. The challenge can come from any citizen – the process varies from state to state, but is codified in the state election law. So yes, typically any voter can challenge someone on the spot. In my state (which is not a voter id state), the Chief Judge at the precinct makes a ruling on the validity of the challenge, usally by calling in to the Board of Election and asking the staff to look some information up. Most of the time it concerns people who have moved but are voting at their old precinct. But if you are in a voter ID state, you either have the id or you don’t and the poll workers will be the ones making the determination toward the provisional.

    Finally, what would keep some group from challenging every voter? There are ‘nuisance’ laws on the books that could be used by a Board of Elections to remove an observer who was disrupting the voting process. It’s a fine line that would need to be walked, but if an observer from one party was willy nilly gumming up the works in a frivolous manner, they could be removed from the polling site by election officials.

    • Matt McIrvin

      In my case, because I presented ID, I got a real ballot. Had I not done so, I believe I would indeed have gotten a provisional ballot instead.

  • Olav Grinde

    MAT, Matt & Zack: Thanks for your descriptions and answers.

    One more question, to you, Ed Freeland, or anyone else: As I understand it, the name on an ID has to match up with a name on a voter roll. However, there has been extensive purging of voter rolls. Some reports claim we are talking about several million unjustly purged voters nationwide.

    Do we have numbers that show how many voters have been purged in each state? In the past? Recently?

    Might this be an even bigger problem than the voter ID issue in terms of skewing elections?

  • William Ockham

    I think the poll results showing almost no Republicans are worried about the voter id laws, combine with stories like Zack’s represent a real danger for Republicans. With Democrats over- concerned about the voter id laws and many paraparty groups deploying compensating programs (like the ads Zack mentions), Democratic may not be affected at all. Concievably, you could have slightly increased turnout.

    On the Republican side, no one is taking the issue seriously. They know it isn’t aimed at them. But some Republicans will be affected. The problem for the Republicans isn’t really in this election, but in future elections. People will be much more likely to give up on voting in disgust if they perceive that their “own side” prevented them from voting.

    Even though I generally prefer Democrats over Republicans, an outcome that led anyone to give up on voting would be a public tragedy that would diminish our democracy.

  • pechmerle

    Well before the election day (though not all early voting dates), there are already court challenges in process to several of these voter i.d. laws passed in the last year or so. I haven’t followed this closely enough, but my impression is that most of the courts — not all — presented with the challenges have been ruling that the i.d. rule cannot be enforced, at least for this election, because it ambushes the potential voter. Some of the challenges that have been ruled on positively were (of course) immediately appealed by a Republican Secretary of State or Governor. Several of those appeals haven’t been ruled on yet, though I would expect the appellate courts to take the timing seriously and move relatively quickly on these.

  • Terry

    Off topic-

    Any of us who watch politicsclosely have been inundated today with a plethora of state polling that seems to show a reduced national vote margin for the President at this point in time.

    In many cases, these individual state ‘snapshots’ are used in support of this subjective conclusion or that.

    If that is not bad enough, Nate Silver has IMO crossed the line toward corporate bloviation by running thousands of simulations to fortell an electoral college tie! Where’s John King and the CNN Magic Board…

    I think that bright people have a right to educate their readers. As for me, 538 is off my favoites list and I’m following the data wherever it goes…

    • wheelers cat

      I returned my copy of the Signal and the Noise today unopened.
      Its against my religion to support global warming denialism.
      Nate Silver has become the enemy of truth in the pursuit of booksales and page clicks.

    • Matt McIrvin

      There could well be some real tightening in the race going on. The past week or two have been unusually favorable for Obama, and the pattern over the previous year has been one of unusual stability.

      The tick down that happened in the PEC EV counts seems to be from some unusually Democratic results from North Carolina getting superseded by more typical ones in the average.

    • Sam Wang

      I agree that based on the entire season, we should expect those indicators to come down soon.

      Note that the EV estimator is noisy because of NC. The Meta-Margin, which fluctuates less, is still moving up. This might be the same as what you’re saying. What it suggests to me is that if there is a peak, state poll aggregation hasn’t hit it yet.

    • Matt McIrvin

      I guess the fundamental question is whether there’s still a hidden setpoint that the race tends to gravitate to (albeit slowly enough that the current high point does matter for November), or if it’s now more like a drunkard’s walk or Markov process in which older history tells us nothing, or something else.

  • Paul K2

    Terry, don’t get too worked up. Nate likely has to do something each day to keep 538 fresh.

    Why not take another look at this NYTimes opinion piece, and especially the comments? I haven’t ever read more informed comments on a political post.

    This debate might actually work like a debate should, and spread some light on real data and forecasts, and if it does, then we can all hope that this year voters actually vote for rationality , intelligence, and competence.

    • wheelers cat

      There is plenty of fresh material for Nate to discuss without helping the GOP elites blow smoke up their base’s ass.
      He just coyly dipped a toe into the “polls lie!” side, incase you missed this.
      Nate: “or that the polls have overestimated his standing across the board. ”
      The title is a lie too.
      “New Polls Raise Chance of Electoral College Tie”
      Its not the polls, its Nate’s initial conditions in his simulation runs. The polls are saying the same thing they have said for a week.

    • Sam Wang

      He’s writing about EV ties? In my view that is really scraping the bottom of the barrel.

    • Sam Wang

      Until I do it of course. (Zinger alert!)

    • wheelers cat

      Scraping the bottom of the barrel for page clicks.
      This new Nate Silver is entirely too interested in book-sales and page clicks for my taste.
      Sure, he was awesome in 2008. This season hes been captured by market forces.
      Its regulatory capture actually, by the conservative cartel. I bet Nates traffic falls off when Romney’s win prob drops below 20%.

    • Olav Grinde

      That’s a very interesting article by David Firestone. Thanks!

    • Matt McIrvin

      People have to write about something when there’s no longer the opportunity for a brokered convention.

  • pechmerle


    According to the Brennan Center for Justice at NYU Law School, voter i.d. law challenges have succeeded, at least in part, in Texas, Florida, Wisconsin, Ohio, and Pennsylvania.

    In the latest ruling, the Pennsylvania Supreme Court sent the law back to the lower court with instructions to consider how the law could be implemented fairly this close to election day. More often than not, that kind of directive from the state’s highest court leads the lower court to take the hint.

  • Iseeurfuture

    PA voter ID law has been halted and will NOT be in effect for NOV 6th.

    • DaveM

      But state officials, having just implemented their “concerted” effort to educate all voters as to the would-be new requirements, are now conveniently out of time to publicize the corrective; hence many voters are likely to be unaware that photo ID is in fact NOT required to vote in this election. How much this may dampen turnout is unknown.

      Beyond that, election officials will ASK for photo ID; they just won’t prevent you from voting if you don’t come up with it. I am planning to refuse their request and will be encouraging every Pennsylvania voter I know to do likewise.

  • Olav Grinde

    Yes, the Pennsylvania court decision is good news.

    I do hope the Democrats are prepared to have at least two proactive poll observers at every single precinct in the country. There is going to be a lot of challenges, intimidations and delays from the Tea Party & Co. Be prepared to see concerted efforts to compel many voters from Democrat-leaning demographic groups to make provisional votes. And, regretfully, I don’t believe all officials involved in the polling and vote count process will exhibit the integrity and neutrality that MAT described above!

    PS. Does anyone have reliable state-by-state figures on how many voters have been purged from the rolls since, say, 2010?

  • Ed Freeland

    Very interesting thread that raises a number of issues:

    First, the impact of provisional ballots: As MAT notes, the margin of victory in most elections will usually exceed the number of provisional ballots. It’s good to know that in close elections where the provisional ballots matter, they are reviewed for validity independent of how the person voted. Also good to know there are lots of lawyers involved.

    The potential impact of voter ID requirements on voter turnout: Zack and William make the case that the voter ID movement could potentially backfire on those who thought it might suppress turnout among their opponents. Greater awareness, preparation or motivation on the part of Democratic voters may counterbalance the potential for any reduced turnout that could result from the new rules. This also speaks to Badni’s point about the number of votes that might potentially be affected.

    On purging the rolls: MAT probably knows more about this than me, but in my experience, large scale data manipulations using name and/or address as the primary identifiers will have lots of random error and so are less likely to affect one party more than another. That, or course, assumes that registered Republicans on the rolls aren’t treated any differently than registered Democrats. Where this really gets wacky is when state election supervisors try to purge non-citizens from the rolls. In Florida, this meant taking information from a federal database and matching it (using name and address) to the merged state voter registration file. Considering all the different ways people can spell their names and addresses, this looks to me like a data nightmare with enough random error to hurt both parties equally.

    On the Enthusiasm Gap: The Quinnipiac University/New York Times/CBS News poll shows that Republicans are consistently more like than Democrats to say they are more enthusiastic about voting in this year’s presidential election than in past presidential elections. By similar margins, Democrats are more likely to say their level of enthusiasm is the same as in past elections. But there’s almost no difference between the two parties in the percent who say they are less enthusiastic about voting this time. In contrast, voters who called themselves Independent are about twice as likely as either Republicans or Democrats to say they are less enthusiastic about voting this year. If there is an enthusiasm gap this year, it’s really between party members and Independents. My guess is that it won’t have much impact on either turnout or the final tally.

  • John L

    I don’t understand the short-sightedness of these laws. The fastest growing voter demographic is Latinos. These are the votes both parties need to be courting for long term survival. These laws have all been placed by Republican legislatures. In future elections all voters will be following the new rules.

    It will be difficult for these same Republicans to gain the trust and vote of the Latino voter that they disenfranchised.

    It may work for one election but the cost could be 10 years of electoral defeat.

  • Jerry Murtagh

    I was thinking that after the election is over and assuming Obama wins, we might address this voter ID issue in a slightly different manner. Let’s agree that we need greater security and move to require states to provide voter registration ID cards that will satisfy the requirements of the state to all currently registered voters and to any new registered voters. Make it the responsibility of the state to issue the cards through registered mail to all currently registered voters. Congress can pass legislation requiring this. I wonder if Republicans in Congress would support such legislation considering their serious concern with voter fraud.

  • Dennis T.

    Thank you for taking up the concerns surrounding Voter I.D. These concerns are real as some of your posters relate. For a more complete, detailed, and highly readable book on this subject see Greg Palast’s brand-new book, “Billionaires and Ballot Bandits.” The whole panoply of voter suppression efforts are covered. The book takes a non-partisan position. this subject took a great deal of time for Greg Palast and associate researchers to investigate over a period of many years. But we do now have a good picture of the whole range of such vote suppression and election tampering schemes, and there are many. I have seen no other book like it. Highly recommended. (Note: The text is not completely “family friendly,” for any parents concerned about what is called profanity. But the content is rock-solid and extremely valuable.)

  • Tushar goyal

    The best advice I could give to members of any political persuasion – if you are worried about what some other group may or may not do on election day – volunteer to be a poll observer with your local party.

  • Voter ID registration

    Voter ID Card registration rules are also changed in India. In earlier days it was done manually, But now a days all procedure gone through website registration.
    Indian govt. also plan to link UID (Unique Identity) with voter ID to get rid of duplicate electors.

    Thnx for sharing such a nice Article.

Leave a Reply to Lindsey K (Cancel)