Although statistics is a rational subject, it does not follow that students can or do learn statistics in a rational way. Students learn statistics in the pattern‐forming mode more than in the rational mode. The paper describes and compares the two modes, and identifies three conditions that prevent or discourage learning in the rational mode. The diagnosis of pattern‐forming de‐emphasizes mathematical derivations and proofs as a means of allowing students to understand statistical concepts. The primary recommendation is that the teacher should deliberately manipulate the pattern‐forming process. The underlying idea is for the teacher to view the process for what it means to the student.