home learning environment (and in particular the home learning environment index9) are
generally higher than for family measures such as mothers’ qualification level. The home
learning environment measures also influence young children’s social/behavioural developmental
gains over the pre-school period. It is interesting to note that the pre-school home learning
environments differ for boys and girls. As a group significantly more girls’ parents reported
activities such as reading, teaching songs and nursery rhymes etc. It is not possible to establish
whether these self-reported differences in parenting reflect different expectations of boys and
girls, and or gender differences in the behaviours and interests of pre-school children. The home
learning environment effect, however, remains significant even when child gender is included in
the models. The results suggest that some of the differences in cognitive and social/behavioural
outcomes at primary school entry evident between boys and girls may in part be attributed to
differences in the quality of home learning environment.
Thus, policies targeted at working with parents in disadvantaged communities (such as Sure
Start) are supported by the EPPE findings. Many pre-school settings across England already
encourage parental participation, and some have developed programmes that feature parent
education. The EPPE results suggest programmes that directly promote activities for parents and
children to engage in together are likely to be most beneficial for young children. Health visitors
may also be well placed to provide guidance for parents on ways to enrich young children’s
home learning environments and some primary schools run activities for parents. Such provision
could also seek to promote the benefits of joint activities, which promote pre-school children’s
developmental learning at home.
Variations in centre effectiveness
Value added multilevel analyses show the individual pre-school centre attended by a child also
has an impact on children’s social/behavioural developmental gains.10 A number of statistically
significant outlier centres were identified (i.e. centres whose children made more progress than
the scores of their children at entry to the study would predict). These are centres where children
showed significantly better (positive outliers) or, by contrast, significantly poorer
social/behavioural developmental gains than predicted (negative outliers), given their prior
social/behaviour and background. There were 52 (36.9%) centres identified as performing
broadly as expected across all areas of social/behavioural development, when intake differences
are controlled. Just over one in 10 centres (12.8%) were found to be statistical outliers
(performing significantly above or below expectation for one or more social/behavioural area).
This is likely to be a conservative estimate of the extent of real differences in effectiveness
between individual centres, since, with small numbers of children per centre an effect has to be
large to reach statistical significance.
Typically centres vary in their effects on different social/behavioural outcomes. No centre
performed significantly above or significantly below expectation for all social/behavioural
outcomes. However, pre-school centre effects are generally more highly correlated in
social/behavioural outcomes than cognitive outcomes. This suggests that pre-school settings
show more internal variation in effectiveness in promoting children’s cognitive outcomes than is
the case for their social/behavioural outcomes. Nonetheless, the most usual profiles across the
four outcomes studied show that a number of centres could be distinguished with broadly
positive effects, whereas others showed generally poorer effects on social/behavioural
developmental gains.
More than a fifth of children (23%) had left their target centre before starting primary school and
moved to other provision. There was no evidence that mobile children, who moved pre-school
centre during the study, showed poorer social/behavioural outcomes when they started school.
9 The home learning environment index provides a summary based on the individual measures reported
above such as parents reading to their child. It is interesting to note that the home learning environment
index is only moderately correlated (r=0.3) with family SES or mother’s qualification levels.
10 Significant centre level variance in children’s social/behavioural developmental gains remains even
when account is taken of prior social/behavioural development and other intake differences (in terms of
child, family and home learning environment characteristics).
30