Did Voter ID Laws Rig the Election?

wisconsin voter ID laws rig election 300,000 votesCourtesy of Dan Arel, with over 47,000 likes and 37,000 retweets.  It was also picked up by many Facebook pages, most notably Occupy Democrats, with over 84,000 likes and 45,000 shares.  This meme claims that 300,000 voters were “turned away” by the Wisconsin voter ID law, thus rigging the election.  On closer examination, this is a highly misleading claim.

Where Does the 300,000 Number Come From?

Certainly not from some independent agency counting the number of voters “turned away” at the polls.  The 300,000 number comes from a US District Court case from 2013, where the Wisconsin voter ID law was being challenged.  The plaintiffs challenging the law brought on an expert to testify that he performed a study finding that 317,735 registered voters in Wisconsin possessed neither a Wisconsin driver’s license nor a state ID card.  Since those are the primary forms of identification voters would use to comply with Wisconsin’s voter ID law, this number has been trotted out as implying over 300,000 voters would be disenfranchised in Wisconsin.  However, there are significant problems with this reasoning, as it relates to the 2016 election.

Errors in Comparing Databases

As Nate Cohn of the New York Times pointed out back in 2014, the real number of registered voters without IDs are much lower than most studies suggest.  In the case of Wisconsin, the 300,000 number was determined by comparing the voter registration database with the DMV database, and seeing which registered voters didn’t have a corresponding license or ID card.  This method is very error prone, as often the two databases won’t match.  People change their names, addresses, have nicknames, etc.  For example, if a teenager got their driver’s license listing their parent’s address, then later registered to vote at their college address, the two databases wouldn’t match.  The study cited in the Wisconsin court case claimed to correct for some of these errors, by trying to find nicknames or hyphenated names for instance, but not enough to correct for all the potential errors.

A Better Study

A better study is found in North Carolina, another swing state.  When they instituted a voter ID law, they had their State Board of Elections perform a 2013 study to see how many voters didn’t have IDs, and how that might affect the election.  When they initially compared the two databases, they found 1.2 million out of 6.4 million total registered voters unmatched with IDs.  After using 29 different criteria methods (seen below) to match names in the two databases (far more than in the WI study), this number whittled down to just 318,000 voters.  It seems it takes much work to find all the names who fall through the cracks.  Considering Wisconsin’s population is just 57% of North Carolina’s, the equivalent number of unmatched voters would likely be around 180,000 in Wisconsin.

North Carolina voter ID study

Table shows how different criteria narrowed down unmatched voters in North Carolina. Source

Voters With Other Forms of ID

Since the Wisconsin voter ID law allows several other forms of ID than a driver’s license or state ID card, it’s very possible that some of those not matching in the database would still have a required ID to vote.  Other allowable documents include passports, military ID, student ID, Indian tribe ID, among others.  For example, an out-of-state student attending UW Madison, or an Army soldier at Fort McCoy, might be a resident of Wisconsin, but never have applied for a new driver’s license, instead keeping the one from their previous state.  An elder member of an Indian tribe might have no need for a driver’s license, but retain his tribal ID.  A repeat DUI offender might have lost their driver’s license, and simply use their passport for ID purposes.  These are just a few example, but each of these people would still be able to provide the valid ID required for voting, yet fall into the “300,000 disenfranchised voters” group.

The opposite could be true, too.  A Wisconsin resident could move to another state or country, and their license could expire, yet their vote registration might still be in Wisconsin.  Or they could die, and still be in the voter registry.  There is simply no way of knowing how many of these voters exist, but considering there’s over 280,000 students alone attending Wisconsin colleges, this could be significant.

Voters Deciding to Obtain ID

The Wisconsin study took place on September 2013, which means it says nothing about how many voters actually didn’t have ID during the presidential election on November 2016.  In fact, the voter ID law didn’t become effective until after April 7, 2015.  It’s very likely some of the voters lacking ID in 2013 decided to get one in order to vote.

Most Without IDs Don’t Vote

This shouldn’t be too surprising.  People who lack the means or initiative to get an ID are far less likely to vote in the first place, registered or not.  The North Carolina study found that out of the 318,000 people who they couldn’t match as having an ID, only 138,000 of them voted in the 2012 presidential election.  That means just 43% of those without IDs voted, compared to the 68.3% turnout of all eligible voters in NC.

Those Without ID’s Wouldn’t All Vote For Clinton

This seems obvious, but many seem to think that the small amount of voters who might be affected by a voter ID law are almost all Democrats.  Again, the North Carolina study shows a different reality.  55% of the unmatched voters were registered Democrats, while 22% were registered Republican.  It’s impossible to know the breakdown of those who don’t vote, but Nate Cohn suggests it’s on the 70/30 level favoring Democratic candidates, not 95/5 like some presume.  Since North Carolina is a close swing state similar to Wisconsin, it would presumably have a similar breakdown.


All of these factors show that the 300,000 number this meme claims will be whittled down substantially.  The following is just an estimate, as there is no reliable data, but it takes far more into account than the meme.

  • Using the more comprehensive North Carolina study and adjusting it for Wisconsin’s population, we’d get 180,000 registered voters without Wisconsin issued driver’s licenses or state ID’s in 2013.
  • Since only 43% of those people were likely to vote, that leaves us with just 77,000 “disenfranchised voters”.
  • Since only 70%, or 54,000 of those were likely to vote Clinton, that leaves us with a net Clinton gain of 31,000 votes.
  • This doesn’t account for all the people who had other forms of ID, left the state, or who decided to obtain an ID in order to vote.  We can’t realistically estimate this, but it would very likely bring the number below 27,000, which was the margin of victory in Wisconsin.

Some could credibly argue with the analysis above and debate whether or not the voter ID law in Wisconsin might have changed the razor-thin outcome there, but it’s wrong to suggest that 300,000 voters were “turned away”, as that’s based on an out of context study with no real relevance to the 2016 election.  Furthermore, it becomes a moot point, as it wouldn’t have mattered if Wisconsin went for Clinton anyway.  Other swing states, without voter ID laws, went for Trump.  Those include Florida, Pennsylvania, and Michigan, where Clinton lost by 1.3%, 1.2% and .3% respectively.  Besides Wisconsin, the other swing states with voter ID laws had such large margins that voter ID laws couldn’t be blamed.  Clinton lost Ohio by 8.6 points, and North Carolina by 3.8 points.

Voter ID laws are a legitimate issue to debate, but to equate them with “rigging” an election is a bit hysterical.  Wisconsin’s law has been challenged, and upheld by the courts.  States have the right to take reasonable steps to prevent voter fraud, and have to weigh that against the reality that voter ID laws might prevent or make it more difficult for some to vote.  It’s possible for a close election to be decided by voter ID laws as well as by voter fraud.  There can be reasonable debate on this subject, but there is no serious debate that any voter ID laws decided the 2016 election.