Growing up in Scotland: children's social, emotional and behavioural characteristics at entry to primary school

Chapter 3 CHARACTERISTICS OF CHILDREN WITH SOCIAL, EMOTIONAL AND BEHAVIOURAL DIFFICULTIES AT AGE OF SCHOOL ENTRY

The aim of this chapter is to identify the children most likely to score in the borderline or abnormal ranges of each of the SDQ scales at entry to primary school. Various background characteristics of children are explored spanning socio-demographic, socio-economic, parenting and early child development factors. Whilst a descriptive overview of the types of children scoring in the borderline or abnormal ranges on each scale is provided, multivariate regression analysis was used to determine which characteristics are related to an increased risk of having difficulties in each of the domains when holding other, potentially confounding, characteristics constant. ⁵

A range of factors were identified that are likely to be associated with social, emotional and behavioural development in young children (Campbell et al., 2000; Shaw et al., 1996). These factors cover the domains of socio-demographic background, socio-economic characteristics as well as specific issues related to early child development and approaches to parenting. For the most part, these characteristics are only measured once during the period of interest, such as ethnicity or mother's age at the child's birth. However, the availability of longitudinal data from GUS allows consideration of changes in key background characteristics which may affect social development such as change in family type or child health. As such, some measures of these changes have been constructed for consideration in the analysis. Appendix 2 presents a detailed description of all factors and explains how they were measured using the GUS data.

Regression analysis allows the association between an explanatory variable (such as income) and an outcome variable (such as severe behavioural difficulties) to be explored whilst controlling for other variables which may affect behavioural difficulties (such as gender). All factors were analysed together using a single binary logistic regression model. The same model was run separately for each SDQ sub-scale and for the overall difficulties scale. The outcome variable in each model indicated whether or not the child's score for the respective scale, when measured at school entry, was above the normal range. ⁶ Thus the analysis identifies which characteristics are statistically significantly related to (or 'associated with') a child having a score in the borderline or abnormal range at school entry. Whilst all variables were included in one model, for clarity the results are presented separately for the different sets of explanatory factors.

3.1 Key findings

• The socio-economic characteristics which were related to difficulties at school entry varied according to the subscale being examined. Overall, these characteristics had limited bearing on behavioural outcomes at school entry. Level of household income was the measure most consistently related to behavioural development with children in lower income households at higher risk of difficulties with conduct, emotional development and hyperactivity than those in higher income households.
• Health and development factors also affected different behavioural domains in different ways. The condition of the child's general health over the period from age 2 to age 5 was consistently and strongly associated with behavioural difficulties at age of school entry. Children with poorer general health tended to have greater behavioural difficulties than those with better health.
• Delays in motor development at age 2 were associated with having more emotional difficulties at school entry. Delays in language development at age 2 were also statistically significant, being associated with greater difficulties in hyperactivity and peer problems as well as a high score on the total difficulties scale.
• Parenting factors, such as use of particular discipline styles or taking children on social visits, tended to be more related to issues of conduct and hyperactivity. Those children who experienced no shouting or smacking, higher levels of parent-child social interaction, and a higher frequency of social visits were less likely to score in the problem ranges of conduct and hyperactivity scales.
• No statistically significant differences in difficulty scores were found according to differences to children's age at school entry.

3.2 Socio-demographic and socio-economic characteristics

Table 3.1 illustrates the proportion of children with particular socio-demographic and socio-economic characteristics who scored in the borderline or abnormal ranges of each of the SDQ difficulties scales and the overall scale at school entry. Those figures which are printed in italics were those which did not remain statistically significant when entered into the regression model. Those which are printed in bold did remain significant.

The socio-economic characteristics which were associated with difficulties at school entry - after controlling for other possible confounding factors - varied according to the domain being examined. On the whole however, very few of these characteristics were associated with scores on the scales when other characteristics were taken into account. Level of annual equivalised income was the factor which was most consistently associated with difficulties at school entry. Although the size of the effect varied dependent on the subscale in question, in general those children living in higher income households had a lower risk of difficult behaviours than did those in lower income households (Table 3.1). For example, one-fifth of children (19%) in the highest income group had scores in the borderline or abnormal ranges for conduct problems at school entry compared with a little over two-fifths of children in the lowest income group (43%).

Living in a house with siblings appeared to reduce the risk of emotional difficulties at school entry. Children with two or more siblings in the household at age 2 were half as likely as children who were the only child in the household to have an emotional symptoms score in the borderline or abnormal range (8% compared with 15%).

Family type and transitions were associated with conduct problems with lone parenting related to greater conduct difficulties. Forty-four per cent of children in stable and repartnered lone parent families had a conduct score in the borderline or abnormal range compared with 31% of children who experienced parental separation and 23% in stable couple families.

Hyperactivity was the only behaviour affected by the child's gender with boys more likely to have difficulties in this domain than girls; 22% of boys returned a score in the borderline or abnormal range compared with 15% of girls. Ethnicity, on the other hand, was the only socio-economic or demographic factor which was associated with peer problems. Non-white children were significantly more likely to have peer difficulties at school entry than were white children (50% compared with 16%). However, the small size of the non-white sample (unweighted base = 56) means that this result should be treated with caution.

Table 3.1 Socio-demographic and socio-economic characteristics of children who score in the borderline or abnormal ranges for each SDQ sub-scale and the total difficulties scale at school entry

	Conduct problems	Emotional symptoms	Hyper-activity	Peer problems	Total difficulties	Weighted	Un-weighted
Sex of child
Male (%)	31	11	22	18	14	1046	1047
Female (%)	26	12	15	15	10	1047	1016
Mother's ethnicity
White (%)	29	11	18	16	12	1985	2005
Non-white (%)	24	17	17	50	16	69	56
Age of mother at birth of cohort child
Under 25 (%)	37	15	26	21	19	503	384
25 - 29 (%)	30	14	21	19	11	468	456
30 - 34 (%)	25	9	13	13	9	654	721
35 or older (%)	22	8	13	13	10	403	476
Number of children in household at age 2
One (%)	26	15	21	19	14	688	663
Two (%)	30	10	17	14	11	926	955
Three or more (%)	30	8	17	19	11	441	445
Family type transitions from age 2 to 5
Stable couple (%)	23	9	15	14	9	1450	1577
Couple who separate (%)	31	22	19	18	14	118	112
Stable lone parent (%)	44	17	28	27	21	316	239
Lone parent who repartnered (%)	44	12	30	21	21	172	135
Equivalised annual household income
1st Quintile (< £11,875) (%)	43	16	30	24	20	474	388
2nd Quintile (>=£11,875<£19,444) (%)	28	14	17	18	14	404	382
3rd Quintile (>=£19,444< £25,625) (%)	25	9	15	13	7	341	360
4th Quintile (>=£25,625< £37,500) (%)	22	9	12	12	8	380	382
Top Quintile (>=£37,500) (%)	19	6	12	12	5	328	382

Note: Figures printed in italics are those which did not remain statistically significant when entered into the regression model. Those which are printed in bold did remain significant.

3.3 Parenting approaches and health and development indicators

Factors associated with parenting experience and child health and development were generally more closely associated with behavioural outcomes at school entry (Table 3.2) than were socio-economic and demographic factors.

The condition of the child's general health over the period from age 2 to age 5 was consistently and strongly associated with behavioural difficulties at age of school entry. Those children whose health was reported as fair, bad or very bad at any point during that time - that is either temporarily (at any single sweep), or always (at all sweeps) - were more likely to have scores in the borderline and abnormal ranges at school entry. For example, 22% of children who temporarily or always had poorer health had scores in the borderline or abnormal range of the emotional symptoms score at school entry compared with 10% of those who reported no poorer health. Indeed, poor health was a very strong predictor of several social, emotional and behavioural issues.

Delays in motor development at age 2 were associated with emotional difficulties at school entry. Seventeen percent of those who had not achieved three or more developmental milestones at age 2 were reported to have emotional difficulties at school entry compared with 5% of those who had achieved all the milestones. Delays in language development at age 2 were associated with difficulties in hyperactivity and peer problems as well as a high score on the total difficulties scale. Notably, early problems with language development was one of the few factors significantly associated with peer problems; 23% of children who were reported to have language difficulties at age 2 had peer problem scores in the borderline or abnormal range at school entry compared with 14% of those with no language difficulties.

Early parenting approaches and experiences were also related to behavioural outcomes at school entry. Use of harsh discipline - that is smacking and shouting - at age 3 was associated with difficulties in conduct and hyperactivity. Children whose parents reported both smacking and shouting were more likely to have difficulties in these domains than children whose parents did neither. For example, 40% of children whose parents used both smacking and shouting had scores in the borderline or abnormal range of the conduct problems scale compared with 19% of children who parents used neither. Frequency of visits to other households with children was the factor most consistenly associated with behavioural outcomes. Those children who had this type of interaction less often than fortnightly, or never, were more likely to score in the borderline or abnormal range for conduct, hyperactivity and peer problems as well as for the total difficulties score. In relation to the peer problems scale for example, 15% of children who experienced this social interaction fortnightly or more often scored in the borderline or abnormal range compared with 23% of children who experienced it less often or never. Low parent-child social interaction was less important, being significantly, but only very weakly, associated with conduct problems and hyperactivity. The direction of the relationship is not clear here. Increased use of harsh discipline, lower levels of parent-child social interaction and reduced visits to other households with children may be responses to already difficult behaviour rather than pre-cursors of it.

Table 3.2 Parenting experiences and health and development characteristics of children who score in the borderline or abnormal ranges for each SDQ sub-scale and the total difficulties scale at primary school entry

	Conduct problems	Emotional symptoms	Hyper-activity	Peer problems	Total difficulties	Weighted	Un-weighted
Child's general health transitions
Always good or very good (%)	26	10	16	16	10	1786	1821
Temporarily or always fair, bad or very bad (%)	48	22	35	24	27	269	241
Delay in motor developmental at age 2
None (%)	25	5	13	15	8	226	223
1 milestone (%)	23	9	17	13	8	371	378
2 milestone (%)	31	11	17	15	9	347	341
3 or more milestones (%)	34	17	23	20	18	479	482
Child's speech difficult for others to understand at age 2
Yes (%)	35	13	25	23	19	1401	1443
No (%)	25	11	15	14	8	655	620
Use of harsh discipline: birth to age 3
None (%)	19	11	14	15	9	461	444
Shouting only (%)	34	13	22	13	13	63	62
Smacking only (%)	26	10	18	16	11	911	924
Shouting and smacking (%)	40	14	21	19	15	621	632
Frequency child taken to visit other people with children at age 2
Fortnightly or more often (%)	27	11	17	15	10	1610	1617
Less often (%)	35	13	24	23	17	442	443
Mean scores on parent-child social interaction scale
Children with scores in normal range	1.41	1.41	1.42	1.42	1.42	1183	1220
Children with scores above normal range	1.35	1.33	1.28	1.29	1.25	136	125

Note: Figures printed in italics are those which did not remain statistically significant when entered into the regression model. Those which are printed in bold did remain significant.

3.4 Differences by age at school entry

As acknowledged above, because of the eligible dates of birth in the original sample design, children in the GUS cohort are spread across two school year groups. Thus to explore characteristics of children in Primary 1 meant combining some data from sweep 3, when around one third of the child cohort had started school, with data from sweep 4, when the remainder had started school. Two key differences exist between these groups which may lead to differences in their scores: first, data collection with those children who started school at sweep 3 occurred, for the most part, in the first school term between August and December whereas that for the sweep 4 group occurred in the spring and summer terms. Second, children who had started school at the sweep 3 interview were younger on average at school entry than were those who had started by their sweep 4 interview, and younger at the time of data collection. Seventy-one percent of children who had started school by the sweep 3 interview were under 5 years old when they started, compared with just 12% of those who had started by sweep 4. However, combining data from both sweeps captures the full spread of ages typical of any single school year group at school entry as well as including data from all school terms.

The appropriate age at which children should start school is a subject which has received much research attention including in relation to its effects on later attainment and social development (Sharp, 2002). To look for any such variations in social, emotional and behavioural development, scores on the SDQ were compared according to the age of the child at entry to primary school.

All local authority schools in Scotland begin the new school year in mid-August whilst some independent schools begin in early September. In order to allow age at entry to be calculated, all children who started school at sweep 3 were allocated a standard start date of 15 ^th August 2007, with those who started at sweep 4 allocated the date 15 ^th August 2008. Figure 3 A displays the distribution of ages at entry to primary one.

Figure 3 A Child's age at school entry

As the graph shows, children are more likely to be aged between 5 and 5 and a half when they start primary one than to be 4 or to be older than 5 and a half. However, there is a reasonable spread of ages with around one-third (33%) starting school when aged between 4 and a half and 5, and a little over one in ten (13%) when aged between 5 and a half and 6.

There were no statistically significant differences in the proportion of children whose scores were classified into each of the normal, borderline and abnormal ranges nor in average scores by age of entry to primary school. Thus the difference in age within a typical school year group does not appear to significantly explain variations in social, emotional and behavioural development at school entry.

3.5 Summary

Factors significantly associated with a score in the borderline or abnormal range for each of the sub-scales and the total difficulties scale at school entry are summarised in Figure 3 B. (see also Table A1.1 in Appendix 1).

The analysis has shown that a range of socio-economic and socio-demographic characteristics, early development issues and parenting experiences act as risk factors for, or protective factors against, the development of social, emotional and behavioural difficulties at primary school entry. Household income and child general health are found, in particular, to be consistently associated with behavioural characteristics at school entry. Across most of the difficulty scales, children living in lower income households and those who experience either temporary or persistently poor health between the ages of 2 and 5 are at increased risk of showing difficulties at school entry.

Figure 3 B Factors significantly associated with a score in the borderline or abnormal range at school

Conduct problems Family type transitions: living in a stable lone parent or repartnered lone parent family. Equivalised household income: living in a lower income household. Child's general health: children whose health was temporarily or always fair, bad or very bad. Discipline: children who experienced shouting and/or smacking at age 3. Parent-child social interaction: children who experienced lower levels of social interaction. Visits to other households with children: children taken on visits less often than fortnightly. Emotional symptoms Number of children in the household: being a lone child. Equivalised household income: living in a lower income household. Child's general health: children whose health was temporarily or always fair, bad or very bad. Delays in early motor development: delayed on 2 or more developmental milestones at age 2. Hyperactivity/inattention Child's gender: being a boy. Maternal age at child's birth: having a younger mother. Equivalised household income: living in a lower income household. Child's general health: children whose health was temporarily or always fair, bad or very bad. Language difficulties at age 2. Discipline: children who experienced shouting and/or smacking between birth and age 3. Parent-child social interaction: children who experienced lower levels of social interaction. Visits to other households with children: children taken on visits less often than fortnightly. Peer problems Mother's ethnicity: children whose mother is non-white. Language difficulties at age 2.