An Inaugural Professorial Lecture This lecture argues that the government's definition of an improving school as being one with a linear, continuous, upward trajectory of test and examination results has passed its sell-by date. There is evidence emerging from longitudinal school improvement studies and from an examination of longitudinal datasets to show that continuous improvement in attainment results from year to year for the vast majority of schools is not a reality. There is also evidence that the use of target setting is having a negative impact on practitioners and, more importantly, on children themselves. The reasons for these findings are complex. Some are in the control of schools, others are not. The implications of these findings for future government policy are identified.