This article is part of Election 2020: America Votes, FP’s 24/7 coverage of the US election results as they come in, with brief dispatches from correspondents and analysts from around the world. The America Votes page is free to all readers.
Since Joe Biden was proclaimed US President-elect on Saturday, outgoing President Donald Trump and his Republican allies have doubled their baseless allegations that the voting process was fraudulent and that Biden was “stealing” the election with “illegal” votes. This is not uncommon in certain countries of the world, where problematic elections are often fraudulent, with or without a basis. Indeed, allegations of election rigging are a classic feature of elections in struggling democracies and authoritarian states.
However, election observers in such countries have opportunities to check the number of votes, which generally makes it possible to detect significant cases of fraud in the aggregation of election results. In this way, they can quickly assess the likelihood of allegations of voting irregularities or fraud – and their possible impact on the election result.
Aside from the fact that there is no basis for Trump’s allegations of systematic fraud, could such techniques have worked in the 2020 US election to assess the possibility of fraud, even if only the allegations were disproved? Unfortunately, highly effective voting verification techniques in other countries either do not work in the US electoral system or have not been tried on a large scale.
Since the 1980s, election observation organizations such as Democracy International, of which I am president, have conducted a form of voting check called Parallel Voting Tabulations (PVTs). PVTs, also known as quick counts, assess the accuracy or verify the integrity of election results as reported by electoral authorities in controversial elections. The way they work is for local election observers to observe the actual voting and census in a sample of polling stations and then independently report the results of the local polling stations. In this way, the monitoring organization can independently assess the accuracy of the aggregated, reported results within statistically significant error limits. While PVTs cannot replace a recount when the margin is paper thin, they generally systematically identify false reports of actual results, the main way in which polls have been rigged in problematic elections around the world.
Could the United States use PVTs to increase confidence in the voting process and prevent frivolous and destructive allegations of electoral fraud from growing in importance? Unfortunately, there are several ways in which elections in the United States currently differ from most other countries that use PVTs. For one, there would have to be a separate, statistically significant PVT for each swing state, as each US state has its own number of votes, rather than having a single nationwide voting system as in most other countries. This increases the complexity and the costs considerably. In general, PVTs are best suited for elections that use paper votes deposited in physical ballot boxes. If elections are conducted using multiple technologies and highly automated voting, as is the case in most US states, PVTs would require vastly different techniques.
In the United States, the national media goes to great lengths to gather enough local information to “name” the election results. As these calls affect public perception, this would indicate the possibility of problems if electoral authorities report different results. In contrast to PVTs, however, these calls are not based on an independent observation of the voting and counting, but on a statistical analysis of the preliminary results of the authorities. Because media projections basically only report the aggregation of counts by local and state authorities, they do not provide much information on whether there were any problems with those local vote counts. And of course, the calls from the media are also being attacked by Trump.
Another way that surveillance organizations around the world often draw conclusions about electoral integrity is to look at exit polls and public opinion polls. In an exit survey, researchers asked selected voters from a sample of polling stations how they just voted. The researchers can then compare the results with the reported results. (Some US media projections also use exit polling as part of their process.) In some cases, international observers have pointed to a divergence between pre-election polls and the results reported by the authorities as the basis for their consultation.
However, the current elections in the United States challenge the accuracy of the pre-election polls, not the other way around. Nobody uses pre-election numbers favoring Biden to indicate that the vote was rigged to favor Trump. Likewise, few would believe that the baseline polls could be used to question the integrity of the reported election results. In addition, a voting or sampling-based process would be too blunt a tool to infer the likelihood of fraud in the census from a very narrow choice, as in most states currently under scrutiny as Georgia and Wisconsin.
To assess the legitimacy of Trump’s attack on the number of votes in the current election, Americans must trust the integrity of state and local election officials and rely on the results of the censuses and potential recounts carried out by representatives from both parties. Fortunately, so far we have no reason to question this process – and every reason to believe that these officials are fully committed to the integrity of the process.