Vote Switching from the Perspective of Data Processing
Posted on 23 Nov 2020; 03:00 AM IST. Last Updated 24 Nov 2020; 07:40 PM IST.Summary: The covid pandemic could have forced some countries to adopt Mail-In voting as a viable alternative to In-Person voting in 2020 elections. This article highlights some of the issues around Mail-In voting, in an election.
The covid pandemic could have forced some countries to adopt Main-In voting as a viable alternative to In-Person voting in 2020 elections.
One of the biggest hurdles of Mail-In voting lies in establishing the Identity of the legal voter. If the identity of the legal voter is not established according to a strict standard, there could be many spurious votes. For example, if drivers license is used as an identification, then almost every resident of a state (as against legal citizens), could vote in the election.
Depending on how they are processed, Mail-In voting could lead to several discrepancies listed below.
a) Allow to cast a vote on behalf of a person, who did not vote.
b) Prevent a person from casting the legal vote.
c) Switch votes between contestants.
d) Result in double counting along with In-Person voting.
e) Invalidate in-Person voting.
f) Allow people who are not legal voters to cast a vote.
The following example demonstrates how Mail-In voting could switch votes between two contestants (here after referred to as “A” and “B”), in some country named “X”.
Assume the following vote counters, while the counting is in progress.
| A | B |
| 10,500 | 9,000 |
If 5000 Mail-In votes arrive, of which 2000 are for A, and 3000 are for B, then we expect the following scenario.
| A | B |
|
10,500 + 2000 = 12,500 |
9,000 + 3,000 = 12,000 |
In such a case, A could be declared the Winner.
But the processing could take a mysterious turn here.
Of the 3000 Mail-In votes of B, if 1000 of them correspond to In-Person votes of A, and if the processing was defined to invalidate previous votes, then we have the following scenario.
| A | B |
|
10,500 + 2000 - 1000 |
9,000 + 3,000 |
In this case, B could be declared the Winner.
This scenario demonstrates that B’s Mail-In votes not only caused an increase to B’s counter, but at the same time decreased A’s counter. This is called “Vote Switching”.
The above scenario, raises several questions.
a) Did the voters who voted In-Person for A, really sent a Mail-In ballot voting for B?
b) If voting allows people to change their vote, then the last vote should prevail. Is the Mail-In really the last choice?
c) Can the 1000 votes, which caused the vote switching, be spurious votes?
It is not easy to declare B as a Winner, when these questions are not answered in good earnest.
From the example(s) given above, it may be noted that spurious Mail-In votes could play havoc, in an election.
The best possible mechanism to prevent spurious votes is to enforce strict “Signature Checks”.
Spurious votes, may be identified by comparing ballots processed by postal services with ballot counts received at counting centers. The example presented below demonstrates the technique.
| Mail-In ballots Requested |
4000 |
| Mail-In ballots Sent |
4000 |
| Mail-In ballots delivered by post office to voters |
4000 |
| Mail-In ballots delivered by post office to counting centers |
4000 |
| Mail-In ballots Received |
5000 |
| Mail-In ballots | 5000 |
The postal evidence demonstrates that only 4000 ballots moved back and forth, whereas the counting centers received 5000.
Conclusion:
Vote switching can also be accomplished by employing intrusion techniques, where an intruder changes the algorithm employed for counting the votes.
Nevertheless, the fraud can be easily detected by a simple hand count or by machine recount with safeguards.
Typically, intrusion techniques play out well in “banana republics”, where the notion of recount may not exist.
In democracies, where recounts and extensive audits are possible, intrusion techniques may be less favoured by fraudsters, compared to spurious votes.
Remarks:
The content of this article and the examples given here do not pertain to any election in any country or state. The data given in the examples are randomly chosen numbers, and any similarity with real world data is purely coincidental.