Establishing a parsimonious model through comparing impact of key student and school factors on secondary school value-added effects

Purpose: – The purpose of this paper is to establish and compare multilevel models that significantly predict school effects on adding value to their students regarding English reading from Secondary One to Secondary Six. Design/methodology/approach: – The data were collected from 3,993 students within 66 schools in 2006 by the Educational Bureau of Hong Kong. Findings: – When entering Secondary One, the students’ English reading ability was tested then again at Secondary Six. A range of value-added models were fitted to the data. The comparison across these models suggests that student individual scores at intake are the most powerful indicators of value-added. The intake aptitude test scores, aggregated to school level, and gender made no significant difference. At the individual level, student band was the significant predictor. School level effects were largely non-significant. Specific findings on value-added across the schools are visualized as evidence of the parsimony of the selected model. Research limitations/implications: – Secondary data such as this while collected at one point in time nevertheless can still shed light on current policies and practices. It is particularly the case considering that the value-added effects system is still working in Hong Kong over decades but less examined academically. Originality/value: – This study has produced some insights for stakeholders to identify influences on the value-added patterns.