The authors (teachers-researcher) carried out an exploratory study in which they designed , on the basis of a didactic analysis similar to the one proposed by Vallecillos (1996), three problems of hypothesis tests. Our purpose was to analyze the effects of graphic calculators use in the application and understanding of the concepts of p-value and significance level in the solving of these problems. The study was done with students of a statistics course for Social Sciences at the university level.<br>We studied, on the basis of the problem solving activities done by the students, their errors and difficulties concerning the concepts of significance lebel and p-value. We thought that the graphic calculator use was going to promote the use of the graphic and numeric representations. However, we found that the graphic calculator was used only in order to represent the p-value numeric representation. Even though each student "seems to understand" the concepts of p-value or significance level because he or she can resolve a problem, we find that little changes in data, (i.e., change in test laterality), generate new errors.<br>This result and other similar ones suggest the need to reflect about the phenomena that put into play the concepts in question, the kind of diactical activities that are designed and used in order to work with these concepts, and the conceptions and obstacles which are behind the errors made by the students.
The CAUSE Research Group is supported in part by a member initiative grant from the American Statistical Association’s Section on Statistics and Data Science Education