gcse

Post date: Nov 13, 2017 9:35:33 PM

We start by analyzing a dataset on inner-London schools that accompanies the MLwiN software (Rasbash et al. 2009) and is part of the data analyzed by Goldstein et al. (1993). At age 16, students took their Graduate Certificate of Secondary Education (GCSE) exams in a number of subjects. A score was derived from the individual exam results. Such scores often form the basis for school comparisons, for instance, to allow parents to choose the best school for their child. However, schools can differ considerably in their intake achievement levels. It may be argued that what should be compared is the “value added”; that is, the difference in mean GCSE score between schools after controlling for the students’ achievement before entering the school. One such measure of prior achievement is the London Reading Test (LRT) taken by these students at age 11. The dataset gcse.dta has the following variables: 

    school: school identifier student: student identifier 

    gcse: Graduate Certificate of Secondary Education (GCSE) score ( score, multiplied by 10) 

    lrt: London Reading Test (LRT) score ( score, multiplied by 10) 

    girl: dummy variable for student being a girl (1: girl; 0: boy) 

    schgend: type of school (1: mixed gender; 2: boys only; 3: girls only) 

One purpose of the analysis is to investigate the relationship between GCSE and LRT and how this relationship varies between schools. The model can then be used to address the question of which schools appear to be most effective, taking prior achievement into account.

Rabe-Hesketh, Sophia; Skrondal, Anders. Multilevel and Longitudinal Modeling Using Stata, Volumes I and II (Kindle Locations 6691-6711). Stata Press. Kindle Edition. 

To read the file in Stata

• • use http://www.stata-press.com/data/mlmus2/gcse, clear