Modelling social, emotional and behavioural development in the first three years of school: what impact do schools have?

Marryat, Louise Jane (2014) Modelling social, emotional and behavioural development in the first three years of school: what impact do schools have? PhD thesis, University of Glasgow.

Full text available as:
[thumbnail of 2014Marryatphd.pdf] PDF
Download (3MB)
Printed Thesis Information:


Social, emotional and behavioural aspects of development are key to children’s overall development. A failure to develop normally in any one of these areas can have far reaching consequences, affecting the child’s ability to learn and to develop relationships with peers, potentially leading to fewer educational qualifications, a lack of future employment, poverty and a range of other outcomes including difficulty forming relationships, mental health issues and increased criminal behaviour (Tremblay et al., 2004; Woodward & Fergusson, 2000; Ttofi, Farrington, & Lasel, 2012; Hodgins, Larm, Ellenbogen, Vitaro, & Tremblay, 2013; Pingault et al., 2013).

In Glasgow City, a large proportion of children live in disadvantaged circumstances, including living in households and areas suffering from multiple deprivation, living with parental substance misuse and witnessing domestic and community violence (Glasgow Centre for Population Health, 2013; Taulbut & Walsh, 2013). These risk factors can all impact on children’s social, emotional and behavioural development (Margolin & Gordis, 2000a; Gennetian, Castells, & Morris, 2010; Chronis et al., 2003). Children also tend to be clustered in schools with other children who may share similar demographic characteristics and who have similar levels of difficulties, which may compound or ameliorate the individual’s strengths or weaknesses.

This thesis aimed to explore the levels of children’s social, emotional and behavioural difficulties at the start of Primary School (age 4-5) and at Primary 3 (age 7-8) in Glasgow city and to investigate the stability of these over time. Analysis was carried out using a brief behavioural screening questionnaire, Goodman’s Strengths and Difficulties Questionnaire (SDQ)(Goodman, 2013b), which had been completed by nursery staff and class teachers. The thesis also aimed to examine whether Glasgow City is different in its levels of social, emotional and behavioural difficulties compared with other areas of Scotland and the UK. Clustering of difficulties within schools is analysed in order to explore the relationships between peer difficulties and demographics, individual difficulties over time and ultimately, whether schools have an effect on the development of social, emotional and behavioural difficulties during the first three years of school.

The results of this study indicate that, between preschool and P3, levels of Conduct Problems, Emotional Symptoms and Hyperactivity/inattention increased, whilst levels of abnormal Pro-social Behaviours decreased and Peer Relationship Problems remained relatively static. Both means and prevalence rates for children in Glasgow City demonstrated similar patterns to UK norms, though levels of Hyperactivity/inattention problems at P3 were higher than in UK 5-10 year olds.

Data from the Growing Up in Scotland study were used to investigate whether a ‘Glasgow Effect’ (i.e. an amount of variation that could not be explained solely by demographic differences in the population) existed in children’s social, emotional and behavioural difficulties at preschool age. Children in the Glasgow sample did have higher rates of social, emotional and behavioural difficulties compared with children in the rest of Scotland. However, this difference in difficulties appeared be entirely accounted for by the difference in demographics within the populations in the different areas. There are various factors which might explain this: sampling issues, such as having to use a Greater Glasgow and Clyde sample rather than Glasgow City, may mask any Glasgow Effect, whilst it may be that differential attrition in the GUS cohort may mean that children with problems are missing from the sample. It could also be that sleeper effects are at work, which may emerge in the form of difficulties later in childhood, or that what we are seeing is a ‘Scottish Effect’ rather than a Glasgow Effect, given that most of the previous research in this area compared Glasgow with demographically similar English cities. At this stage however, it appears that results from Glasgow may be generalisable to other areas, once demographics are controlled for.

Multilevel modelling of Strengths and Difficulties Questionnaire (SDQ) data from Glasgow City schools was then used to explore what factors were associated with longitudinal increases in SDQ scores between preschool and P3. Results showed significant differences between schools in the unadjusted models, accounting for 11% of variance in change scores. The adjusted model found that having worsening social, emotional and behavioural difficulties in the first three years of school was associated with being male, being from a White-UK background, and having had Looked After status (been under the supervision of the state) by preschool. Being in a school with a small school roll was also associated with an increase in difficulties over this time. School effects were only reduced slightly in this final adjusted model, accounting for 9% of variance between schools, suggesting that variation in the development of social, emotional and behavioural difficulties may differ somewhat between schools during the first three years. It should be noted that numbers of pupils within schools were small in some cases, leading to wide confidence intervals and possibly reducing significance of the results.

Having social, emotional or behavioural problems at P3 (above the cut-off on the SDQ for likely difficulties) was also related to a range of factors. Again in the unadjusted model, there was a statistically significant difference in levels of difficulties between schools. The strongest predictor of such problems was having had an abnormal score at preschool. Also important was being male, having been Looked After by preschool and being in a school with a higher proportion of children eligible for free school meals, which is likely to be a proxy for income deprivation. However, once these characteristics were controlled for, there was no statistically significant difference between schools.

The ability to identify a group of children prior to the start of school who are at risk of continued social, emotional and behaviour difficulties raises questions about whether a preschool mental health screening test should be put in place. It would be hoped that this would allow children to access the support they need in order to optimise their development, with a suggestion that a universal screen for these types of problems could double or treble the traditionally low numbers of children receiving help (Goodman et al., 2000). However, there are also difficulties with a screening tool of this nature, including the potential for false identification of difficulties, the potentially negative impact of labelling children, the additional burden that this may place of services and finally the current lack of evidence around a potential effective intervention for children of this age (Goodman et al., 2000; Sayal et al., 2010; Wichstrom et al., 2012).

In conclusion, children in Glasgow City have similar prevalence rates of social, emotional and behavioural difficulties at preschool, compared with children in the rest of the UK, but these difficulties are markedly worse in Glasgow City by the third year of school. However, the difference in these scores may just be due to Glasgow City having a more disadvantaged population, which in turn impacts on levels of difficulties. Whilst schools were found to make a difference in relation to children moving up or down the spectrum of difficulties in the first three years of school, there was no evidence that schools contributed to the likelihood of children having an ‘abnormal’ score at P3, though again, this lack of significance could be related to the small numbers of pupils within some schools, as well as the over-riding impact of having difficulties earlier in life.

More research is required with larger numbers of students within schools in order to see if the lack of variance seen between schools at P3 is real or whether it is a sampling issue related to small numbers and therefore wide confidence intervals within schools. Qualitative work around the outliers and some of the unusual findings, e.g. that children in smaller schools appear to fare worse, would be beneficial in interpreting the findings. It would also be of great benefit to follow these children up to the next stage of data collection at P6, in order to explore what happens to children’s social, emotional and behavioural difficulties by the end of Primary school.

Item Type: Thesis (PhD)
Qualification Level: Doctoral
Keywords: Child development, preschool, school, behavior.
Subjects: L Education > L Education (General)
R Medicine > RJ Pediatrics
Colleges/Schools: College of Medical Veterinary and Life Sciences > School of Health & Wellbeing > Mental Health and Wellbeing
Supervisor's Name: Thompson, Dr Lucy, Minnis, Professor Helen and Wilson, Professor P.
Date of Award: 2014
Depositing User: Dr Louise Marryat
Unique ID: glathesis:2014-6302
Copyright: Copyright of this thesis is held by the author.
Date Deposited: 11 May 2015 08:25
Last Modified: 02 May 2018 12:22

Actions (login required)

View Item View Item


Downloads per month over past year