The aim of the study is to analyse the middle school students’ estimation performances (the accuracy of estimation) in regard to measurement based on several variables, including their attitudes towards estimation, their experience in regard to estimation, gender and grade level. The study employs the quantitative methods, namely causal comparison and relational designs. The participants of the study are 490 middle school students attending public schools in Ankara. Data collection tools are Measurement Estimation Skills Test, Estimation Attitude Scale and Estimation Experience Scale. The findings of the study indicate that the estimation performance of most of the students is moderate. It is also found that the students had lower success in the area estimation in contrast to the length estimation. While there is no statistically significant difference among the variables based on the gender, it is found significant difference according to grade level. While there is a statistically significant relationship between the participants’ estimation performance and their attitudes towards estimation, a similar correlation is not found with estimation-related experience.