Debunking Elections

Proving that something has happened is often quite hard, because no evidence was left behind, or because the evidence that was left has been lost or destroyed. Proving that something did not happen is even harder, because if nothing happens, no evidence is left behind. The focus has to be on what was done to prevent something happening.

The United States has monitored elections in countries all around the world, condemning many as stolen and fraudulent, often with very little evidence. Debunking an election is quite easy. Political commentators just throw out some accusations, regardless of whether they have evidence, and force the administrators of the election to refute their claims. They often just repeat the claims of the opposition politicians who lost the election without bothering to check if they are true.

Proving these claims of fraud are wrong is quite difficult, because there is usually no evidence to demonstrate that nothing happened. And when an accusation is refuted, the political commentators just throw out some different ones. Their strategy is often to make so many complaints about the conduct of the election that the election administrators are kept on the back foot. They are often left repeating the claim that the election was secure. However, the more frequently they repeat this claim, the less plausible it sounds. If they shout the message that the election was secure too loud, it begins to sound ludicrous.

The US had debunked elections all over the world, usually with very little evidence of fraud, but mostly because the incumbent does not submit to the demands of the sitting US president. So it is ironic that the latest US presidential election is being challenged as fraudulent and there is does not seem to be any way of proving conclusively that it is fair. I can’t help thinking about people with planks in their eye looking for specs in the eyes of others.

I listened to Chris Krebs the ex-director of the Cybersecurity and Infrastructure Security Agency (CISA) explaining why the presidential election was safe. His comments mostly felt into the category of repeating the message that the election was the most secure ever. He did not seem to have much evidence that this statement was true.

The main direction of his efforts seems to have been ensuring that paper ballots were used. He correctly explained that this reduces the risk of one type of fraud (flipping votes on tabulating computers) because votes can be recounted by hand. The problem is that using paper ballots opens the door for other types of fraud, and he did not mention what he had done to ensure that these types of fraud did not occur. Postal ballots further increase the risk.

I have been involved in projects that involved scanning paper records and tabulating the results on computers. There are standard ways of checking that the process is secure. I would have expected Krebs to be able to describe the things that were done to ensure that the entire process was secure.

• Audits of the process for sending postal voting forms to voters.

• Audits of the security of the process for delivering postal votes from postal to election centers.

• Audits of the process of batching, recording, moving and storing paper ballots within the election centers to ensure that all votes can be accounted for at all times. This requires sound administration practices, which are well understood.

• Audits of the counting process to ensure that there was no scope for smuggling fraudulent paper ballots into the election centers.

• Independent checks of the design and construction of the voting and tabulating computer hardware to ensure that are what their manufactures claimed them to be.

• Checks that voting machines were not vulnerable to cyber-attacks.

• Testing the machines in advance to ensure that the software actually does what it was expected to do. Software testing is a very well developed profession

• Auditing the design of access to computers to ensure that human error does not affect the way that they function.

• Monitoring the process for setting the scanners to ensure that they are reading and recording accurately.

• Auditing the process for aggregating the data from different voting machines to ensure that there was no scope for data manipulation or data loss. It seems bizarre that this was done in some places by transporting USB drives from place to place.

A well-prepared quality plan would have identified in advance, all the opportunities where fraud could occur, and put processes in place to ensure that they did not happen. These are standard quality management practices that are well understood. I am surprised that the election officials are not publishing their quality plans and the audits done to ensure that actual practices complied with their quality design.

I wonder if this stuff was not done as well as it might have been because the administration of the election is so decentralised, with each state having different processes. In some states, counties seem to have managed the election process differently. This variation in practice makes it difficult to do enough of the quality management processes described above. Assessing all variations would be too difficult.