From Toby Pereira, Rayne, Essex, UK
Bob Holmes describes a poll revealing that most people said a serial killer was more likely to be a teacher who didn't believe in god than a teacher (19 August, p 22). It seems strange that the study combined a test for anti-atheist bias with a logic test: the correct answer was that the murderer was a teacher, because the set of teachers is larger than those of atheist or believer teachers.
The way the experiment was set up, people's prejudice – either for or against atheists – would only come out if they first failed the logic test. It seems less surprising that those who fail a logic test would also have an irrational bias.
