How Trajectories of Disadvantage Help Explain School Attainment

This article illustrates the links between different ways of assessing disadvantage at school and subsequent qualification outcomes at age 16 in England. Our previous work has compared variables that represent current or recent snapshots of disadvantage (such as eligibility for free school meals [FSM]) with long-term summary variables and found the latter to improve measures of both social segregation between schools and explanations of raw-score differences in attainment. This new work takes an even more detailed longitudinal approach, modeling the course of one age cohort of 550,000 pupils from the National Pupil Database through their entire schooling to the age of 16 in 29 distinct analytical steps, using “effect†sizes, correlations, and a regression model. The steps represent stages such as what is known about each pupil when they were born, who they attended school with at age 10, and where they lived at age 14. The model also includes variables representing where data are missing for any pupil in any year. Using capped Key Stage 4 points as an outcome measure, these stages can predict the outcome with R = .90. This is considerably higher than for models using either snapshots or summaries of disadvantage. Key predictors are poverty and special educational needs at age 5, and throughout schooling, coupled with prior attainment at ages 6, 10, and 13. With predictors fed into the model in life order, there is little evidence of differential progress for different language and ethnic minority groups and no evidence of regional differences or a type of school effect. The article concludes with the implications of these results for assessing disadvantage when considering school contexts and for policy makers. Given the small but apparently consistent negative school composition “effects†in every year, one clear implication is that school intakes should be as mixed as possible both socially and academically.