3. Establishing a representative sample of PIPS data for Scotland and a stable sample of schools in Scotland for trend analysis
Data were analysed from three cohorts of pupils in Scotland; those children in Primary 1 during the 2012/13, 2013/14 and 2014/15 academic years. The 2012/13 cohort was analysed in most detail. This particular cohort has previously been analysed and its representativeness established (Tymms et al., 2014). The procedures are described in the Tymms et al. (2014) research report but for the reader's ease, they are also described in Appendix A of this report.
For the trend analysis, the same sampling process was used for the 2013/14 and 2014/15 data as was used for the 2012/13 data. The Scottish Index of Multiple Deprivation (SIMD) deciles and the percentage of boys/girls were used to establish samples that were representative of Scotland and so were comparable across the three years. Data were excluded if an assessment date was missing or was unrealistic, if the intake was recorded as Term 2 or 3, or if one or both total scores were missing.
Exactly the same numbers of pupils were taken in each decile as were used in the 2012/13 sample.
3.1 Characteristics of the 2012/13 sample
3.1.1 Sample size
A total of 6,627 children were included in the Scottish sample for the 2012/13 academic year. These children all had start of Primary 1 data for the cognitive development part of the assessment. The assessment of personal and social development was optional and not all children were assessed at the start and end of the year. The numbers in the sample are detailed in Figure 1.
|Cognitive development||Personal & Social development|
|Start of Primary 1||6,627||669|
|End of Primary 1||6,626||396|
3.1.2 Home background (Scottish Index of Multiple Deprivation)
The proportion of pupils in each SIMD decile nationally in Primary 1 is mirrored in the sample for the 2012/13 cohort as Figure 2 shows.
|Number in sample||788||722||656||656||656||656||656||656||656||525||6,627|
|Proportion in sample||0.12||0.11||0.10||0.10||0.10||0.10||0.10||0.10||0.10||0.08|
|Number in P1 2012||6,841||6,146||5,568||5,593||5,456||5,417||5,736||5,567||5,440||4,585||56,349|
|Proportion in P1 2012||0.12||0.11||0.10||0.10||0.10||0.10||0.10||0.10||0.10||0.08|
The age of children starting school in Scotland is shown in Figure 3.
The 2012/13 data set contained dates of birth for 6,615 pupils which enabled us to calculate their age at the start of school. The mean age of the pupils starting school in Scotland was 5.03 years. Figure 3 illustrates the spread of ages at the start of school and it is evident that a number of children started school when they were five and a half or older. This was also reported in Tymms et al. (2004) and mirrors trends seen in the Growing Up in Scotland study, which published findings in 2012 indicating that 9% of children were older than five and a half years when they started school in the cohorts that they had studied. Interestingly, 15% of boys had their entry deferred compared with 9% of girls in the study (Bradshaw et al., 2012). In Scotland, parents can request to defer their child's entry to school to the following year under certain conditions and this leads to a number of children who are older than might be expected at the start of Primary 1. It is clear from Figure 3 that although the numbers are low, this approach was not restricted to very young children; based on the score linked to ages it is clear that those with delayed development are more likely to be kept back for a year. This is explored in Appendix E.
Children start school, on average, when they are six months older than children starting school in England. The distribution of ages at the start of school in England is more rectangular: For every month of birth there was about the same number of children. Although in England a small proportion of children start after Christmas and an even smaller proportion start after Easter.
The 2012/13 data set included 49.8% females and 50.2% males.
Figure 4: Proportions of males and females in the 2012/13 samples of PIPS data and the national data
|PIPS sample Scotland||49.8||50.2|
|PIPS sample England||48.1||51.9|
|National data for Primary 1 in Scotland 2012||48.9||51.1|
The 2011 Census in Scotland reported that slightly more boys are born than girls (National Statistics 2012) with 51.2% of males in the 0-14 year's population. This is very close to the proportions of males seen in the national data for children in Primary 1 in 2012. The PIPS sample in Scotland has a slightly lower proportion of males to females, however, the proportions do vary by region, as reported by indexmundi (http://www.indexmundi.com/facts/united-kingdom/quick-facts/scotland/sex-ratio#map).
3.1.5 Pre-school experience
Although teachers were asked to complete information about their pupils' pre-school experience as part of the PIPS project, it is clear that many teachers did not complete that section of the PIPS questions; they may not have had that information to hand. Summary statistics are published by the Scottish Government (2014). The number of children in early learning and child care centres each year is reported but no information about the amount or quality of provision. The numbers are reasonably stable but there is a slight decline over time between 2007 and 2010, then an increase up to 2014.
3.1.6 English as an Additional Language
As with pre-school experience, the information provided by teachers within the PIPS assessment was sparse but the Scottish Government's Education Analytical Services Division extracted aggregated data relating to the schools included in this study as reported in Figure 5.
Figure 5 indicates small and slightly increasing proportions of children from ethnic minority groups and children whose first language was not English in the sample. This is close to the pattern across Scotland as a whole.
3.2 Characteristics of the data sets for Primary 1 2012/13, 2013/14, 2014/15 trends over time analyses
3.2.1 Trends over time
The total sample of the three cohorts analysed to investigate trends over time comprised 19,578 pupils from 394 schools. The mean number of pupils in the schools was 50 (Standard Deviation =34.4; range between 1 and 168; median of 44). Details of the three cohorts are given in Figure 6. The figures for the 2012/13 cohort in the trends analysis are slightly different from the 2012/13 data used for the baseline and follow up analysis because the analysis excluded a number of additional cases as noted above.
The SIMD profiles, ages and proportion of boys/girls were consistent over time.
|Number of pupils with data||6526||6526||6526||19578|
|Male, n (%)||3263 (50.4)||3345 (50.9)||3410 (51.2)||10018 (51.0)|
|Female, n (%)||3218 (49.7)||3177 (49.1)||3116 (48.8)||9511 (49.0)|
|Mean (SD)||5.0 (0.31)||5.0 (0.31)||5.0 (0.31)||5.0 (0.31)|
|Median (min, max)||5 (4.5, 6.3)||5 (4.5, 6.5)||5 (4.5, 6.4)||5 (4.5, 6.5)|
|1, n (%)||734 (11.3)||734 (11.3)||734 (11.3)||2202 (11.3)|
|2, n (%)||717 (11.0)||717 (11.0)||717 (11.0)||2151 (11.0)|
|3, n (%)||651 (10.0)||651 (10.0)||651 (10.0)||1953 (10.0)|
|4, n (%)||653 (10.0)||653 (10.0)||653 (10.0)||1959 (10.0)|
|5, n (%)||653 (10.0)||653 (10.0)||653 (10.0)||1959 (10.0)|
|6, n (%)||651 (10.0)||651 (10.0)||651 (10.0)||1953 (10.0)|
|7, n (%)||650 (10.0)||650 (10.0)||650 (10.0)||1950 (10.0)|
|8, n (%)||650 (10.0)||650 (10.0)||650 (10.0)||1950 (10.0)|
|9, n (%)||644 (9.9)||644 (9.9)||644 (9.9)||1932 (9.9)|
|10, n (%)||523 (8.0)||523 (8.0)||523 (8.0)||1569 (8.0)|
|Missing||0 (0.0)||0 (0.0)||0 (0.0)||0 (0.0)|
Email: Wendy van der Neut