Do any voting system models try to take "realistic" tactical voting into account?

Question

In standard voting system models (see example), there are two types of voters: honest voters who always choose their preferred candidate(s) and tactical voters who try to maximize the expected weight of their vote. However, in reality, voters are neither perfectly honest nor perfectly tactical; they may not rank more than one candidate on a ranked ballot due to lack of understanding of the system or lack of interest in researching other candidates.

Are there any voting system models that take this into account?

Answer 1

Kenneth Arrow proved, in his famous impossibility theorem, no system can be guaranteed to give a reasonable outcome on several criteria all the time:

In short, the theorem states that no rank-order electoral system can be designed that always satisfies these three "fairness" criteria:

• If every voter prefers alternative X over alternative Y, then the group prefers X over Y.
• If every voter's preference between X and Y remains unchanged, then the group's preference between X and Y will also remain unchanged (even if voters' preferences between other pairs like X and Z, Y and Z, or Z and W change)
• There is no "dictator": no single voter possesses the power to always determine the group's preference.

But that doesn't mean that some practical systems are not better than others or mostly work well.

But there are specific undesirable outcomes promoted by first past the post systems. Like tactical voting. In the UK, for example, many Labour (left wing) voters may vote LiberalDemocrat (centrist) to out a Conservative (right wing) candidate which some may think is undesirable as it blurs the indicator of what voters actually want in a government.

Single Transferable Vote systems in multi member constituencies (as used in Ireland and some elections in Northern Ireland) vastly reduce the incentive for voters to misstate their "true" preferences as the selection process for winners does a good job processing the ranked preferences in aggregate (not perfect as Arrow showed, but far better than in an FPTP system).

Practical experience shows that voters do not have big problems understanding such systems and that, in practice, they are better at representing their preferences than FPTP alternatives.

Answer 2

This is sort of a frame challenge so I hope it still answers your question.

If you try to design a voting system or compare different voting systems to each other you generally want a voting system that produces the results that would correspond to 100% honest voters.

An individual voter however wants their own candidate to win. So given a voting system and a voter who wants a particular candidate to win, one can ask how this voter should vote to maximize the chances of their candidate. A voter doing this is a tactical voter. Hence what you want in an ideal voting system is that a tactical voter behaves exactly the same way as an honest voter.

To me the conclusion here is that if you only look at voting systems abstractly it is not meaningful to consider more cases than just pure honest and pure tactical voters.

However this changes if you look at real life voters in a real life voting system. Here you can definitely ask whether actual voters are more honest or more tactical, whether this mixture depends on the candidate they support or even whether there are failing tactical voters who try to vote tactically but end up doing worse than an honest vote. I am not aware of any studies trying to look into these kind of questions, it also seems very difficult to judge these things in a secret election.