For the well-functioning of a direct democracy it is essential that citizens have access to and know about the relevant facts when they are called to vote on policy proposals. In direct-democratic campaigns, however, both proponents and opponents of a policy proposal have incentives to present only selected facts that serve their own position and help sway public opinion in the desired direction. In this chapter, the authors propose a novel typology of how statistics – one of the most often used tools to substantiate arguments with seemingly objective facts – can be (mis)used in direct-democratic campaigns. The typology is based on the distinction between misinformation and disinformation. Whereas the former represents unintentionally false information, the latter refers to the deliberate distribution of inaccurate information. The authors identify four major types of misuse of statistical information: (1) flawed statistics, (2) false predictions, (3) misleading examples, and (4) manipulated numbers. They illustrate each type with an example drawn from recent direct-democratic votes in Switzerland. The examples show that mis- and disinformation in vote campaigns can have far-reaching political and legal consequences.