- The only non-score peer effect in Cohort 1 with a statistically significant effect on the intercept is pupil turnover.
- Pupil turnover, in the operationalisation adopted here, was highly statistically significant in both cohorts: no other ‘non-score’ peer effect were statistically significant after allowing for pupil turnover.
- Pupil turnover had a negative effect in both cohorts, and a possible mechanisms would be the disproportionate time the teacher had to spend inducting and socializing the incomers.
- Peer group effects show to be stronger at older ages.
- The measures of deprivation (proportion free school meals, proportion free school clothes) have a considerably weaker effect, if any.
- Findings show that the other pupils in a class do have an impact. It is more difficult for children to make progress in some classes than others. Class composition should be taken into account in assessing value added.