Can someone in a household coerce a remote voter?

Yes. When a person votes away from a supervised polling place, someone in the same room can watch, instruct, or take the device, and encryption does nothing to stop that. This is one of the strongest objections to any remote voting design, including anything My Vote Forward might describe. The framework treats coercion resistance as an unresolved requirement for any future official remote vote, not as a solved feature, and the websites conduct no official votes.

The objection in its strongest form

The objection holds that the secret ballot is not mainly a technology. It is a place: a booth, supervised by officials, that a voter enters alone. The booth makes it impossible for a spouse, parent, partner, landlord, employer, or community leader to see the ballot being marked, and so makes it pointless to demand a particular vote. Remote voting removes the booth. Whatever happens on the device, a coercer standing beside the voter can see the choice, dictate it, or mark it.

The strongest form adds three points. First, the people most exposed to household coercion, such as those in controlling relationships, dependent older adults, or people in tightly knit communities, are also the least able to report it. Second, coercion leaves no technical trace, so audits that check whether ballots were counted correctly cannot detect it. Third, the harm scales quietly: a coercer does not need to control many voters to matter in a close local contest, and no one would ever know.

Why the objection matters

The objection matters because remote participation is part of what makes ongoing civic engagement possible for people who cannot easily reach a polling place. If remote channels cannot protect free choice, the framework would be offering convenience at the cost of the very privacy its organizing language promises: “My vote.” It also matters for priorities and opinions, which the framework already collects remotely in its demo. Those are not votes and carry no official weight, but the same household can pressure a person’s answers.

Defined terms

• Coercion: pressure, threats, or control that makes a person vote or answer other than as they freely would.
• Family voting: the practice by which one member of a household supervises or directs the votes of others.
• Coercion resistance: a property of a voting design that makes coercion ineffective, for example by letting a coerced vote be replaced later.
• Revoting: allowing a voter to cast a new vote that replaces an earlier one, so that only the last counts.
• Supervised voting: voting in a polling place where officials control the voting area.

Evidence

International standards name family voting as a violation. The Venice Commission’s Code of Good Practice in Electoral Matters states: “Voting must be individual. Family voting and any other form of control by one voter over the vote of another must be prohibited”.⁠¹ Its explanatory report describes family voting, “whereby one member of a given family can supervise the votes cast by the other members,” as something that “infringes the secrecy of the ballot” and calls it “a common violation of the electoral law”.⁠¹ The Code adds that “voters should always have the possibility of voting in a polling station,” and that postal and electronic voting are acceptable only under conditions, including that for postal voting “fraud and intimidation must not be possible”.⁠¹ It notes the “heightened risk of ‘family voting’” in postal voting.⁠¹

Expert review recommended against internet return of ballots. In 2018, a National Academies of Sciences, Engineering, and Medicine committee concluded that “Internet voting should not be used at the present time” and that “Ballots that have been marked by voters should not be returned over the Internet or any network connected to it”.⁠² The report’s reasoning covered security as well as secrecy; coercion is one of several concerns, not the only one.

Federal risk guidance, United States. In 2020, CISA, the EAC, the FBI, and NIST issued guidance on electronic ballot delivery, marking, and return, originally for official use only and reissued publicly in 2024 as guidance that “remains operative”.⁠³ Verified Voting reports that the agencies concluded electronic ballot return is “high-risk” even with safeguards in place.⁠⁴

Some laws define coercion broadly. The United Kingdom’s Elections Act 2022 defines undue influence as a corrupt practice that includes “using or threatening to use violence against a person,” “causing or threatening to cause financial loss,” “causing spiritual injury to, or placing undue spiritual pressure on, a person,” and “doing any other act designed to intimidate a person,” when done to compel a person to vote a particular way or not to vote.⁠⁵ A law can punish coercion after the fact; it cannot see into a living room.

How comparable systems have handled it

Estonia: revoting and a paper override. Estonia’s internet voting lets a voter vote electronically as many times as they wish; “Only the last i-vote cast is taken into account, and earlier votes are annulled”.⁠⁶ A paper vote at a polling place also overrides an i-vote; the State Electoral Office states that from 2021 voters can change their i-vote by voting on paper at a polling place on election day.⁠⁶ The office explains the purpose as ensuring “that the voter expresses their actual will,” including where a voter “finds that they were influenced during the casting of i-vote”.⁠⁶ Its reliability page adds that voters can choose a suitable time and place to vote, that buying i-votes “is a crime,” and that “Transfer of ID-card and its codes is prohibited”.⁠⁷

Revoting is a real countermeasure, and it has limits that follow from its design: it helps only a voter who has a later private moment and the freedom to use it, and a coercer who controls the credential or the voter’s time can prevent that. Estonia’s own guidance relies on law enforcement for vote buying,⁠⁷ which is an acknowledgment that the technical design does not stop it.

Postal voting: legal conditions and limited eligibility. The Venice Commission suggests postal voting may be confined to people in hospital, imprisoned, with reduced mobility, or residing abroad, and allowed only where the postal service is safe and reliable.⁠¹ Practice varies: the Code’s explanatory report notes that postal voting “may be widespread in one country and prohibited in another owing to the danger of fraud”.⁠¹ The coercion objection applies to any ballot marked outside a supervised booth, paper or electronic.

Competing interpretations

One interpretation holds that remote voting’s benefits for access outweigh its coercion risk, especially for voters with disabilities, voters abroad, and voters who cannot take time off work. On this view, the right response is legal penalties, revoting, and an in-person option, as in Estonia,⁠^6,7 rather than refusing remote channels.

A second interpretation holds that coercion risk is unmeasurable and therefore unbounded. Because coercion leaves no trace, no study can show it is rare, and a system that cannot measure its own failure should not be used for official decisions. The National Academies’ 2018 recommendation is consistent with caution on remote internet voting, though it rested on several grounds.⁠²

A third interpretation distinguishes stakes. Coercion of a non-binding priority ranking is a harm to the person and to data quality, but not to an election outcome. On this view, remote collection of priorities and opinions can be acceptable with clear labels, while remote official voting faces a much higher bar.

What the framework answers

My Vote Forward’s rules are:

• Remote official voting is a research problem. Any future official remote channel would need jurisdiction-specific legal authority, independent evaluation, and an answer to household and employer coercion. Nothing in the demo implements it (see intimidation).
• An in-person alternative remains essential. Any remote channel would be an addition to supervised voting, not a replacement (see intimidation).
• There is no proof of a ballot: no share button, no vote history, no “I voted for” card, which removes one tool a coercer could demand (see private voting).
• Priorities and opinions can be revised at any time, and only the person’s current answer counts in later results (see what if someone changes position?).

Where the objection still wins

The objection wins. The framework has no design that prevents a person in the room from watching or directing a remote vote, and it knows of no remote design that fully does. Revoting helps some voters some of the time. An in-person option helps only voters who can reach it. Removing proof of a ballot helps against coercers who demand proof, not against those who simply watch. For priorities and opinions in the demo, the same pressure can shape answers, and the framework cannot detect it.

Unresolved questions

• Can any remote design make coercion ineffective rather than merely punishable?
• How would a future system measure coercion, given that it leaves no technical trace?
• Should remote official voting, if ever authorized, be limited to voters who cannot reach a polling place, as the Venice Commission suggests for postal voting?
• How should coercion of non-binding priorities be weighed against their value as participation?

Related

• Intimidation
• Private voting
• Secret ballot
• Does a fingerprint prove eligibility?
• What happens when the network fails?
• Hard questions