In an election, self-voting, i.e. candidates voting for themselves or their own proposals, might only capture an obvious inclination or a fear of loss of reputation, and hence may not be useful towards choosing the best candidate. In some contexts, e.g. for small scale boardroom elections, it can thus be sensible to prohibit self-voting, especially, this will prevent everybody pointing to themselves as the best choice. In the case of public elections this is easy to enforce, however, in standard secret ballot elections the no-self-voting condition is unchecked and relies on the honesty of the participants. More commonly, the constraint is simply not imposed in the first place due to lack of enforcement. A generalisation is where certain groups are not allowed to vote for their own candidate. In this case, preventing self-voting can also reduce the level of coercion, e.g., if team leaders demand, or more subtly simply expect, all their team members to vote for them in an election for the best team leader. With the aid of secure e-voting, imposing the no-self-voting constraint becomes possible. We show how this constraint can be implemented efficiently, in the context of both centralised and decentralised voting. Especially, we show how to obtain a robust (i.e. allowing absentees) decentralised voting system preventing self-voting by using just standard linkable ring signatures and anonymous vote-casting channels.