C. Palop Melgar1, M.M. Lopez-Martin2, P. Arteaga Cezón2

1I.E.S Zaframagón de Olvera (SPAIN)
2University of Granada (SPAIN)
Our study focuses on the study of correlation, a concept that is studied in the first course of high school (the level of participants in our sample). The teaching of correlation is framed within the study of two-dimensional distributions and the relationships between two statistical variables. This concept together with regression have great relevance in the domain of Statistics. At a first level, these concepts are widely used in statistical research processes, where knowledge about the intensity of the relationship (correlation analysis) between the variables involved offers enough information to mathematize such a relationship (regression analysis) in the fitting of a model for predictive purposes. At a second level, they are the basis for understanding other advanced statistical concepts and procedures.

The study of correlation has great relevance in the training of high school students and university courses, since it increases their knowledges of one-dimensional descriptive statistics. Sánchez Cobo, Estepa and Batanero (2000) describe five different representations to work correlation and regression: verbal description, frequency table, scatter plot and coefficient of correlation. In this work we focus the analysis of the last activity “scatter plot”. This type of activity gives us the opportunity to delve into the study of correlation and at the same time allows us to develop fundamental skills and abilities, such as the ability to plan a process or to analyze and interpret results.

The sample is composed of 32 students in the subject Mathematics I (Sciences specialty) and 22 students in the subject Mathematics Applied to Social Sciences (Humanities and Social Sciences specialty). The results show that 50% students of Science and 30% of Humanities and Social Science have correctly completed the activity. On the other hand, we emphasize that 34.73% of Science students and 50% of the other modality have had the correct answer, but they have not been able to justify it. The results obtained in this activity are in line with the research carried out by Sánchez-Cobo (1998), where the author points out the existence of errors in the estimation of the correlation from the scatter plot.