ABSTRACT: Proceeds from the abundance of statistical information type present in everyday life has originated an interest in statistical training, particularly at the level of students in initial teacher training. Consistent with the above, the purpose of this article is to describe the domain of decoding skills in information in problematic statements and graphical representations using measures of central tendency. For this purpose, a comparative evaluation of inferential cross-section was performed, using a stratified sample of 73 students of Pedagogy in Mathematics from two universities of the Council of Rectors of the center-south, the first legal status of public and private second. An instrument duly validated by expert and statistically throwing a Cronbach's alpha α = 0.88 and content validity index of 0.71 was applied. In the numerical analysis of information descriptive methods (arithmetic mean, standard deviation and frequency percentage) and inferential methods (comparison test of proportions) they were used. The results show that in general, students in Initial Teacher Training (FID) have descended achievement rates in relation to the calculation of numerical summary measures, such as average weighted arithmetic average or when the statements are implicit character, as well as with regard to the simulation data distributions under (known average) initial conditions. From the above it can be inferred that students usually do not provide statistical theoretical knowledge in using measures of central tendency in relation to the nature of the variables.
Key words:
Measures of central tendency.Teacher initial preparation; Statistical skills assessment.