This is a chapter in a book titled The Myth of the Rational Voter: Why Democracies Choose Bad Policies. As you can guess, the chapters preceding this one detail the apparent irrationality of voters.
This chapter seeks to address an objection. Since people are not generally irrational, why think that they become irrational when they vote?
Caplan uses a simple economic model to explain why it would make sense for people who are generally rational to behave irrationally when voting. Roughly speaking, being wrong is costless to them. In fact, it can often be beneficial.
By contrast, the incentives are reversed when we make decisions about what to buy and where to work.
At the end of the chapter, Caplan contrasts his view with one presented by Brennan and Lomasky. They interpret voters as acting expressively rather than instrumentally. Voters act as they do because they know their votes will not make a difference; that explains their apparently irrational behavior. Caplan, by contrast, maintains that voters actually believe in the policies and politicians they support. They are not just expressing their values.
I wanted to read this because we have had a lot of interest in misinformation in general and voting in particular. I think this is a useful framework for analyzing the phenomena that people are interested in. If we start with the observation that acquiring knowledge can be costly, I think we can get a handle on why misinformation is a problem.