This article presents the results of a qualitative research with a group of 15 university students of social sciences on informal inferential reasoning developed in a computer environment on concepts involved in the confidence intervals. The results indicate that students developed a correct reasoning about sampling variability and visualized reasonable intervals of variability in a repeated sampling, at the same time students identified correct relationships between sample size and confidence level in the width of an interval and margin of error, and identified the randomness of a confidence interval. However, they had difficulties conceptualizing the confidence level as the percentage of intervals that capture the parameter in a sampling repeated under identical conditions. [Written in both Spanish and English. For the complete proceedings, see ED583608.]