A news story claiming that the cream filling of a popular "double" cream sandwich cookie is not really "double" went viral in fall 2013. A high school mathematics teacher posted a blog entry describing how he and his students measured 20 cookies, analyzed the data, and concluded that the double cream cookies had only 1.86 times the filling of the regular variety (Anderson 2013). The report was both encouraging and troubling to a group of mathematics educators. They were excited to see real-world applications play a prominent role in school classrooms. However, investigative reports highlighting the use of very small, non-representative samples and dubious data collection strategies are disturbing; they reinforce statistical misconceptions and misunderstandings. Mathematics educators developed a project that encouraged students to implement statistical experiments of their own design while confronting misunderstandings perpetuated by the news story. This article describes that project--the Cookie Conundrum--through displays and discussions of students' work. Although the Cookie Conundrum was conceived as one for use with undergraduates or pre-service teachers, the project could certainly be adapted for use with mathematics students in the middle school or secondary school grades.