Fighting Fear of Fraud: What Individuals Can Do to Restore Confidence in US Elections
Elections are not merely a way to choose winners. They are also a way to make losers believe they have lost. When losers accept their defeat, they strengthen the democratic process by thinking ahead to the next election, deciding what they will do differently. When they fail to accept their defeat, they weaken the democratic process by challenging the legitimacy of the past election. In the worst case, civil unrest ensues. Elected officials may be assassinated, government buildings blown up, coups plotted.
Because accepting defeat requires soul searching within a party about what went wrong, there’s always a temptation to unify around “we were robbed.” In the past, candidates have found it easier to resist that temptation than rank and file party members have. That’s because political pros are more familiar with the precautionary checks and balances built into the mechanics of election administration. The average citizen understands casting a vote about as well as turning on a light switch. That is, they understand their own role at the surface of the process, but not what the election administrators (or electric utilities) do behind the scenes.
Unfortunately, 2016 is different. Three months before there are any general election results, one of the major-party presidential candidates is already stoking distrust of those results rather than tamping it down. As a result, ordinary citizens will have to assume responsibility for sober assessment of the electoral process. Here are four suggestions for what the ordinary person can do to calm the fear of election fraud.
Use Precise Language and Demand It From Others
Unsubstantiated claims of election fraud will come across as plausible if they connect with a broader, bipartisan narrative of “rigged elections.” To prevent that narrative from becoming the accepted conventional wisdom, use precise language to describe elements of the political process that you see as in need of reform.
For example, you might appropriately suggest that a particular party eliminate super-delegates from its nominating process, or that it use primaries rather than caucuses. Or you could advocate for the adoption of ranked choice (instant runoff) voting in general elections so that voters can more safely express a preference for a third-party candidate.
Each of those topics, and many others, is fair game for debate. Stay specific, though. Make sure what you are advocating is clear. Don’t obscure it with vague terms for the actors (“they” or “the establishment”), activities (“the system”), and defects (“rigged” or “broken”).
To shape the tenor of conventional wisdom, it isn’t enough to be precise in your own language. Whenever someone at the water cooler or bar gripes about “the system” being “rigged,” politely press them to be more specific.
Watch Election Administrators at Work
Learn how election officials ensure that eligible voters can vote, ineligible individuals can’t, and the votes are accurately counted and reported. Not that these processes are perfect — those who work in this area are constantly looking for refinements and opportunities to import best practices from one jurisdiction to another. But they are taking a lot of care that you probably don’t know about.
Although you can read about election procedures, you’ll gain a better appreciation if you watch some of them in action. You’ll be able to more credibly push back against fear of fraud if you speak as a first-person witness. The opportunities for observation vary from state to state; what follows are two examples from Minnesota, one on election day, the other afterward. These two processes provide important protection against fraud and also protect against non-malicious errors and equipment failures. Yet few Minnesotans even know these protective processes exist, let alone have exercised their right to observe them.
When the election judges close a precinct polling place at the end of election day, they don’t simply print out a results tape from the ballot scanner. Among other things, they count the number of voters who signed in and verify that this count matches the number of ballots cast. They also ensure that the number of blank ballots the precinct received matches the total of ballots cast, ballots set aside as spoiled or duplicated, and ballots remaining unused.
These steps detect any situation where someone inserted extra ballots into the ballot box or, conversely, removed some of the ballots. So long as the ballot box remained locked, neither of those things ought to have been possible, but the whole point of audit processes is to provide an extra, publicly observable cross-check.
You can show up at poll closing time and watch this painstaking process play out. Minnesota election law specifically grants the general public the right to observe poll closing activities. Be respectful, though, of the election judges, who are engaged in a stressful process that requires exacting attention to detail. If they lose count, they have to start over, extending an already long day. Therefore, please observe from a reasonable distance and don’t interrupt with questions about the process.
What about the risk that the ballot counters might not have counted the ballots correctly? Could malicious software have deliberately altered the counts? Various precautions guard against this, some as simple as restricting physical access to the ballot counters, so that it is difficult to install modified software. However, you don’t need to trust the programmers, technicians, or guards — there’s a simple, direct cross-check that you can observe.
After the election is over, each county randomly selects a subset of its precincts for hand counting. This “post-election review” is not the same as a recount: it is done regardless how close the election is and regardless of whether a candidate requests it. You can find out the location, day, and time your county has chosen for its post-election review; it will be somewhere between the 11th and 18th day after the election.
On that day, the general public is again legally permitted to observe from a respectful distance. You will watch bipartisan teams of election judges manually examining each ballot, sorting them into piles, and counting the number in each pile. They will ensure that they have accurately counted the ballots and they will compare their counts against those that were printed on election night by the machines. That way, you don’t have to trust all the technical details of the machines’ operation: you can see with your own eyes that the machine counts are confirmed by humans .
Observing these hand counts is not riveting. Beyond the tedium of the process, the outcome is underwhelming. Perhaps one or two ballots will be identified where the voter did not properly fill in the ovals. Any adjustments to the vote counts will be very minor. But that’s precisely the point: you can see that election administrators are willing to perform a great deal of labor just to show that there isn’t any big discrepancy. If you observe that care, you can respond to cynics who seem to think election administrators are easy marks, just blindly waiting to be defrauded.
Expect to be Surprised
When the actual weather differs from the forecast, nobody thinks the thermometers or rain gauges are rigged. And the survey professionals who forecast elections (or explain them afterward, using exit polling) uniformly take the election results as the gold standard against which they judge their work, not vice versa. They understand that their surveys can be misleading for any number of reasons. In particular, they may not understand how the survey respondents compare with the voters, or they may not understand how those individuals’ survey responses compare with their actual votes.
Election results that differ from exit polls are not a smoking gun. Nor are election results that differ from expectations in other ways. For example, some people have an idea about the statistical patterns they expect to see when comparing large precincts with small ones. Upon looking at election data from many precincts, they find that these patterns don’t actually arise. Repeatedly, and across wide areas, the actual data fails to fit the expectation. The logical conclusion is that the expectation is based on an incomplete understanding of how voters in different sized precincts behave, not that fraud is rampant, yet slipping undetected past audits such as the post-election review.
The one time when surprising results may actually be a sign of trouble is in the unofficial results released on election night and in its immediate aftermath. Seasoned election professionals and campaign operatives know to keep an eye out for precincts that are way out of their normal range. For example, a usually lopsided precinct may come in with exactly equal numbers of votes cast for two candidates. That’s a red flag that in the rush to report preliminary results, a number got transcribed from the wrong column.
Errors in the election night reporting can be unnerving. For days after the election, the totals will keep changing as more and more of the errors are found and fixed. The key thing to keep in mind, though, is that the unofficial election night reporting does not feed into the official results, even after corrections. The official results are tabulated much more carefully by well-rested clerks working under less time pressure. Until that process is complete, the updated election night reporting gives you a better approximation to what is coming than exit polls do. However, you should still expect some surprises.
Keep Voter Identification in Perspective
Every state conducts its elections differently. One of the highest profile areas of difference is how voters are identified. In Minnesota, the voters identify themselves by stating their name and address. In some other states, they need to show an official document. This makes it harder to vote, whether one is a fraudulent impersonator or simply a voter without the requisite documentation at hand. However, this distinction does not mean that states such as Minnesota can expect significant fraud through the appearance of impersonators at their polling places.
The 2016 election is not the first with strongly motivated participants. People have poured billions of dollars into trying to win previous elections. Nor is it the first election in which many states allow undocumented voters — the recent court cases have generally preserved the status quo. Yet there have been close to zero incidents detected in which someone signed into a polling place under a false identity.
Ah, but just because impersonation isn’t detected, does that mean it isn’t occurring? The fearful seize upon this question. They would have you believe that without careful checking of identity documents, there would be no way to detect an impersonation. If that were true, then a lack of detections would be meaningless. But it isn’t true.
An impersonation might be detected in any number of ways. Some of those ways might result in the perpetrator being caught. For example, the staff at the sign-in table might personally know the target of the impersonation. Or the impersonator might brag about their crime to someone who rats on them. Other forms of detection might let the perpetrator get away. For example, the target of the impersonation might show up to vote and find their line in the poll book already signed. Or election officials removing a deceased individual from the election rolls might find a record of that individual having voted between the date of death and the date of removal.
None of these things are happening. Not only aren’t impersonators getting caught, they also aren’t leaving tell-tale traces behind. Election administrators are looking for all those things, and they aren’t finding them. Some advocacy groups claim that they have found such signs themselves. They have turned in lists of hundreds of suspected cases. But these have invariably turned out to result from the advocacy groups not checking as carefully as the election administrators do. For example, a suspiciously signed line in the poll book turns out to be a simple off-by-one error. Or a case of John Doe voting after death turns out to be John Doe, Jr., voting after John Doe, Sr., had died.
Could an impersonator get lucky and avoid detection? Sure, in any individual instance. However, throwing even a close election would require hundreds or thousands of impersonations. The math is daunting. Suppose you are such a clever fraudster that you can impersonate someone with a 99% chance of going undetected. If you attempt two impersonations, the chance of getting away with both of them — not being detected in either — is .99 squared. Three impersonations would be .99 cubed. And 300 impersonations would be .99 raised to the 300th power, which happens to be less than 5%. That is, even if each time you are nearly sure to go undetected (99% sure), once you’ve committed the crime 300 times, you are nearly sure (95% sure) of being detected somewhere along the way.
Putting that math together with the lack of detected impersonations gives strong evidence that nobody is trying to rig an election by committing hundreds of impersonations. Also, keep in mind that nobody knows in advance just how close an election is going to be. Senator Franken was elected in 2008 by a 312-vote margin, but if anyone had tried to rig that election, they wouldn’t have known in advance that 313 impersonations was enough. Even though it would have been a recklessly large crime spree, it wouldn’t have been enough to provide any confidence of victory. Perhaps that’s why even strongly motivated campaigns concentrate on turning out legitimate voters instead.