The aim of this contribution is to present an approach to item analysis in a self-examination questionnaire for students in Statistical Data Analysis for Psychology. We have worked with a sample of 413 students from the Faculty of Psychology from the University of Barcelona during the 2008-09 academic year. They answered three different closed questionnaires (20 to 30 items each) about general and specific statistics topics. The study of the results was made according to the item response theory. The third questionnaire (tools of inference statistics) was the best one in discrimination rates, while the second one (probability models) presents a balanced distribution of items in terms of their rates of difficulty. Finally, the first questionnaire seems to be the easiest one (descriptive statistics). The conclusions show that self-examination can be a useful system of student assessment but not for the entire topic in a course on statistics for psychology students.