This thesis discusses how assessment data are used to support the learning of pupils aged four to seven years in four Kent primary schools. The sample was 451 pupils in Reception, Year 1 and Year 2; the researcher collected and analysed quantitative data from pupil attainment on school entry – either as Baseline Assessment or the Foundation Stage Profile – and from results in reading, writing and mathematics at the end of KS1, both as SATs and Teacher Assessment. These data were triangulated with qualitative data collected from a semi-structured questionnaire, classroom observations and interviews with the Reception class practitioners. The author – a head teacher with many years’ experience of primary schools and the Early Years – outlines recent and current government policy and links these to assessment theory and existing practice in the four schools studied. She identifies some possible influences on attainment and looks at how value-added data are currently used as measures of pupil performance. The three research questions look at whether benchmark data can be used to predict future achievement, the educational implications of using value-added data as measures of pupil performance, and whether benchmark data can be used to support learning in the primary classroom. The findings led the researcher to conclude that accurate prediction from prior attainment is not possible at the present time and that contextual value-added data are only useful when other variables are taken into account. However, the findings showed that benchmark data – when used formatively – can be useful in supporting pupils’ learning. This study will help head teacher colleagues to look at data in a fresh way, and to identify and target the needs of individual pupils to optimise their performance from the beginning of the Foundation Stage to the end of Key Stage One.