Investigation of the middle school students' statistical thinking

The purpose of this study was to investigate middle school students? statistical thinking across the 4 process according to SOLO taxonomy. In addition, it was aimed at determining how students? statisrical thinking levels change with respect to the grade level, gender, and mathematics achievement.This study was conducted at an elementary school in the center of Bolu during the fall semester of 2000-2009 academic year. Participants were 30 students attending 6th, 7th and 8th grades at this elementary school. Ten students from each grade level were selected according to maximum diversity sampling. This selection was done by their mathematics teachers according to their mathematics achievement level, which was categorized as low, normal, and high. Data about students? statistical thinking were obtained with semi- structured interviews, and from student?s solutions and drawings during the interviews. Seven problems measuring the processes of describing, organizing, representing, analyzing and interpreting data were used. Collected data were analyzed using content analysis technique.Results of the study indicate that students progress through 4 levels of thinking within each statistical process. These levels of thinking were consistent with the cognitive levels in the SOLO taxonomy. Results also show that there is a consistency in students? statistical thinking across the four statistical processes. Five students had levels of thinking that were consistent across all four processes. Eighteen students had levels of thinking that were the same for three processes, with the fourth process one level higher or lower. Five students had the same level of thinking for two processes, with thinking for two other processes rated one level lower. Only two students? levels of thinking spanned three consecutive levels of statistical thinking across the four processes.The results of the study also showed that students? statistical thinking levels were higher in describing data in compared to other 3 processes. Moreover, students? statistical thinking levels showed no significant difference with respect to the grade level. Also, male students' statistical thinking levels were better than female students? levels. This can be attributed to the fact that male students? mathematics achievements was better than female students?. In addition, students' with high level of mathematics achievement accumulated at upper levels of thinking whereas students with low level of mathematics achievement accumulated at lower levels of thinking.Key Words: Statistical Thinking, statistical four process, SOLO taxonomy, elementary education