Growing up in Scotland: the impact of children's early activities on cognitive development

Chapter 4 Factors Influencing Cognitive Development

4.1 Key findings

• Cognitive ability was measured in children at age 34 months using assessments of their language development (naming vocabulary) and problem solving skills (picture similarities).
• Large variations in cognitive scores were evident at age 34 months with children from less advantaged families outperformed by their more affluent counterparts on both assessments.
• Children who had been identified at 22 months as having developmental difficulties had lower cognitive ability scores than children with no developmental difficulties. Children born with low birth weight and boys also had lower than average ability scores.
• Children whose mothers have no qualifications scored less well than those with degree-educated mothers (who perform particularly well, especially on the naming vocabulary assessment). Children with older mothers (30 years or above) perform better than those with younger mothers.
• Children in households with four or more children have lower ability scores than those with fewer or no siblings. Low household income levels and unemployment/low working hours are also significant factors associated with poorer performance.
• Those in the 15% most deprived areas of Scotland have lower scores than those in the rest of Scotland.
• Children who were read to often, and those who had visited a library by the time they were 10 months old, scored higher on both assessments than children who had comparably less experience of these activities (though the impact was less pronounced for the picture similarities assessment than for the naming vocabulary).
• At age 22 months the number of days in the past week children had played educational games, their overall daily activity levels and the number of places or events they had visited in the past year were all associated with cognitive ability. The more activities children had experienced the higher their ability scores.
• Ability scores were higher among children whose parents rated four or five activities as very important, and whose parents were satisfied with the range of their activities, than for children whose parents attached less importance to activities or were dissatisfied with their range.
• The association between activity levels and cognitive ability might simply be a reflection of the fact that children with high activity levels tend to be from more socially advantaged backgrounds; this is explored further in the next chapter.

4.2 Introduction

This chapter looks at variations in cognitive ability scores across a number of different socio-demographic factors as well as the activities measures outlined in the previous chapter. The intention here is to present an overview of interesting patterns of association based on a selection of factors, rather than to provide an exhaustive exploration of cognitive ability. All of the factors presented below will be considered further in the next chapter where the question of how these all inter-relate will be addressed.

4.3 Socio-demographic factors

The socio-demographic factors considered in this analysis can be classified into four types:

• Child level factors:

• Birth weight
• Gender
• Developmental difficulties identified at 22 months

• Mother level factors:

• Age at birth
• Education level

• Household level factors:

• Number of children in household
• Equivalised income
• Family and employment type

• Area level:

• Scottish index of multiple deprivation

Not only are the above factors strongly associated with activities, as shown in Chapter 3, they are also strongly related to each other. This is explored in the next chapter.

The most striking point to note from all the tables presented below is the extent to which large variations in ability are evident from a very early point in life. At just 34 months of age children from less advantaged families are outperformed by their more affluent counterparts on both the measures (naming vocabulary and picture similarities). Although individual differences in performance are shaped by a vast and complex range of factors, including many genetic or biological factors which are beyond the scope of being captured in a study like GUS, there is compelling evidence that social environment is a hugely critical determinant of outcomes. The implications of this in terms of appropriate policy interventions, and the likely impact that these early developmental differences might have on these children's outcomes later in life, are not insignificant.

4.3.1 Child level factors

Table 4.1 presents the mean T-scores ³ for both ability measures and how they vary by birth weight, gender and whether the child was identified as having some developmental difficulties at age 22 months. This latter measure is based on the Infant-Toddler Checklist component of the Communication and Symbolic Behaviour Scales ( CSBS) ⁴ that was administered at sweep 2. This identifies children whose communication development is of concern, for example that they are not using the kinds of language or gestures expected of a child of their age. The strength of association with both cognitive ability scores is notable - children identified at 22 months as being of concern have much lower scores than children for whom no concerns were identified. This suggests that the CSBS assessment has been a useful tool within GUS for identifying children at risk of later developmental problems. Although they constitute a relatively small group within the population overall, they are an interesting group whose later outcomes are worth monitoring as they age. In addition, as the number of cognitive and developmental assessments carried out in GUS increases over time it might be possible to identify children whose developmental indicators suggest persistent concern and look at the risk factors associated with this status.

The finding that ability scores are lower for boys and those with low birth weights is consistent with evidence from previous studies (as already discussed in section 2.4).

Table 4.1 Average cognitive ability T-scores at age 34 months by child level factors

Socio- demographic factors	Mean score	Bases
	Naming vocabulary**	Weighted	Unweighted
Birth weight
Low	48.1	256	233
Not low	51.9	3594	3648
Gender
Boy	49.3	1990	1995
Girl	54.0	1928	1954
Developmental difficulties at 22 months
Some difficulties identified	44.0	262	248
No difficulties	52.5	3284	3341
	Picture similarities**
Birth weight
Low	46.5	256	233
Not low	49.9	3593	3646
Gender
Boy	48.6	1990	1994
Girl	50.8	1927	1953
Developmental difficulties at 22 months
Some difficulties identified	45.4	262	248
No difficulties	50.1	3284	3339

**Differences significant at less than .001

4.3.2 Mother level factors

Turning now to look at factors related to the children's mothers shows that both age and education are associated with ability scores. Children whose mothers have no qualifications scored less well than those with degree-educated mothers who appeared to perform particularly well, especially in respect of the naming vocabulary scores. Having an older mother (30 years or above) also appeared to confer an advantage.

Table 4.2 Average cognitive ability T-scores at age 34 months by mother level factors

Socio- demographic factors	Mean score	Bases
	Naming vocabulary**	Weighted	Unweighted
Mother's highest educational qualification
Degree	55.5	1091	1245
Vocational below degree	51.2	1489	1513
Higher grade	51.3	306	315
Standard grade	49.6	688	602
No qualifications	45.6	332	263
Mother's age at birth
Under 20	47.5	294	202
20 to 29	49.9	1582	1461
30 to 39	53.6	1888	2117
40+	54.4	128	147
	Picture similarities**
Mother's highest educational	qualification
Degree	52.1	1090	1244
Vocational below degree	49.8	1487	1512
Higher grade	50.0	306	315
Standard grade	47.9	688	602
No qualifications	44.8	332	263
Mother's age at birth
Under 20	45.9	294	202
20 to 29	48.9	1582	1461
30 to 39	51.0	1887	2115
40+	49.9	128	147

**Differences significant at less than .001

4.3.3 Household level factors

The household level factors include a measure of income that has been adjusted to take account of the household size, employment patterns (which can also be a proxy measure of income), and a measure of the total number of children in the family. This latter measure will in part reflect the way in which financial resources are spread within a household, with costs rising as family size increases, but it is also a useful measure of the possible distribution of time resources within a family. One possible hypothesis is that children in large families spend proportionately less time interacting with their parents than children in smaller units. Alternatively, it is possible that the greater opportunity for sibling interactions in larger households confers a developmental advantage on children in such families. Table 4.3 suggests that children in households with four or more children have lower ability scores than those with fewer or no siblings, and that low income levels and unemployment/low working hours are also significant factors associated with poor performance. However, larger family sizes are also commonly associated with lower income households so the analysis will need to be taken a little further to untangle the patterns evident here. The next chapter addresses this.

Table 4.3 Average cognitive ability T-scores at age 34 months by household level factors

Socio- demographic factors	Mean score	Bases
	Naming vocabulary**	Weighted	Unweighted
Number of children in household
One	53.2	1326	1289
Two	51.9	1764	1841
Three	49.2	631	637
Four or more	46.4	198	182
Equivalised household income quintile
Up to £8,409	46.0	871	719
£8,410-13,749	50.5	785	761
£13,750-21,784	53.0	685	723
£21,785-33,570	54.6	742	820
£33,571 and above	56.0	604	693
Income unknown	51.8	231	233
Family and employment type
Lone parent working >16 hours	51.1	304	272
Lone parent unemployed/working <16 hours	45.6	440	339
Couple both working >16 hours	54.3	1712	1863
Couple one person working >16 hours	51.1	1291	1330
Couple both unemployed/working <16 hours	45.2	161	134
Number of children in household
One	50.3	1325	1289
Two	50.2	1763	1839
Three	48.1	631	637
Four or more	46.8	198	182
Equivalised household income quintile
Up to £8,409	45.5	870	718
£8,410-13,749	49.4	785	761
£13,750-21,784	50.3	685	722
£21,785-33,570	51.9	741	819
£33,571 and above	52.7	605	699
Income unknown	50.0	231	233
Family and employment type
Lone parent working >16 hours	50.2	304	272
Lone parent unemployed/working <16 hours	44.3	439	338
Couple both working >16 hours	51.2	1712	1863
Couple one person working >16 hours	50.1	1290	1329
Couple both unemployed/working <16 hours	44.7	161	134

**Differences significant at less than .001

4.3.4 Area deprivation

This final socio-demographic indicator uses a neighbourhood level measure of deprivation so goes beyond the confines of individual or household circumstances. Table 4.4 compares the scores of children living in the 15% most deprived of areas in Scotland with those in the rest of the country. For both assessments children in the 15% most deprived areas have lower scores than those in the rest of Scotland.

Table 4.4 Average cognitive ability T-scores at age 34 months by Scottish Index of Multiple Deprivation

Socio-demographic factors	Mean score	Bases
	Naming vocabulary**	Weighted	Unweighted
Scottish Index of Multiple Deprivation
15% most deprived areas	48.2	662	534
Rest of Scotland	52.3	3257	3415
	Picture similarities**
Scottish Index of Multiple Deprivation
15% most deprived areas	46.0	660	533
Rest of Scotland	50.5	3256	3414

**Differences significant at less than .001

4.4 Children's activities

The reasoning behind the selection of the following activities is set out in Chapter 3 and will not be repeated here. However, a reminder of the measures being explored is useful:

• Activities at age 10 months

• Frequency of reading with the child
• Whether the child ever visits the library

• Activities at age 22 months

• Number of days in the past week the child had played educational games (recognising shapes, numbers, letters)
• Overall daily activity levels in past week (composite measure based on 5 activities, in quartiles)
• Number of different places/events child had been to in past year

• Parental attitudes at 22 months

• Parent's perceptions of the importance of playing educational games
• Number of activities parents rated as very important for their child to do
• Parental satisfaction levels with child's range of activities

As was the case with the socio-demographic factors discussed above, the patterns of association between activities, parental views and cognitive ability scores are striking. However, a similar - or indeed stronger - note of caution needs to be flagged here in terms of what these results do and do not mean. While it is very evident that ability scores increase in line with rising activity levels, there is not sufficient proof that one causes the other. The results presented below are only part of the investigation into why children's outcomes differ so markedly, the discussion in the next chapter expands on this and goes further by identifying the influence these activity measures have once other socio-demographic factors have been taken into consideration. This section should not, therefore, be read in isolation without reference to the analysis still to come. With this in mind the following commentary simply highlights the key findings from the tables and does not go into extensive detail about their possible implications.

4.4.1 Activities at 10 months

Both measures of early 'literary' activity show an association with ability scores, with children who were read to often, and those who had visited a library by the time they were 10 months old, scoring higher on both assessments than children who had comparably less experience of these activities. Although statistically significant, the absolute size of the difference between the picture similarity scores for children who had been to a library and those who had not is actually quite small so its overall impact is probably marginal. This type of finding is common in studies with large samples such as GUS.

Table 4.5 Average cognitive ability T-scores at age 34 months by activities at 10 months

Activities at 10 months	Mean score	Bases
	Naming vocabulary**	Weighted	Unweighted
Frequency of reading/looking at books with child
Every day	53.0	2560	2653
Once or twice a week	50.5	790	772
Once every one or two months or less	47.7	277	258
Never	45.1	225	201
Whether child ever visits library
Has been	53.5	1003	1079
Has never been	51.0	2848	2803
	Picture similarities**
Frequency of reading/looking at books with child***
Every day	50.5	2558	2653
Once or twice a week	48.6	790	772
Once every one or two months or less	48.1	277	258
Never	45.5	225	201
Whether child ever visits library
Has been	50.6	1002	1077
Has never been	49.3	2848	2803

**Differences significant at less than .001

4.4.2 Activities at 22 months

All three activity measures at 22 months show significant associations with ability scores and the magnitude of the differences between groups is also quite notable. For example, the mean naming vocabulary t-score for the 25% most active children (based on their activities in the previous week) is 56.0 compared with 45.1 for children in the 25% least active group.

Table 4.6 Average cognitive ability T-scores at age 34 months by activities at 22 months

Activities at 22 months	Mean score	Bases
	Naming vocabulary**	Weighted	Unweighted
No. of days child played educational games in past	week
Every day	53.8	1138	1149
Four-six days	53.6	528	541
One to three days	51.5	971	986
None	48.9	1204	1196
Overall daily activity level quartile
First quartile (most active 25%)	56.0	895	932
Second quartile	53.1	1112	1136
Third quartile	51.2	1068	1093
Fourth quartile (least active 25%)	45.1	760	706
Number of different places/events visited in past year
Five or more	55.3	725	814
Four	53.9	799	849
Three	51.5	983	993
Two	50.0	840	790
None/one	45.9	507	438
No. of days child played educational games in past week
Every day	50.7	1139	1150
Four-six days	51.0	528	541
One to three days	49.5	970	985
None	48.3	1202	1194
Overall daily activity level quartile
First quartile (most active 25%)	52.1	896	933
Second quartile	50.8	1110	1134
Third quartile	49.3	1067	1092
Fourth quartile (least active 25%)	45.9	760	706
Number of different places/events visited in past year
Five or more	52.1	723	812
Four	51.2	800	850
Three	49.8	983	993
Two	48.1	838	789
None/one	46.1	507	438

**Differences significant at less than .001

4.4.3 Parental attitudes towards, and satisfaction with, activities at 22 months

Table 4.7 looks at parent's assessments of the importance of activities and their satisfaction with their children's range of activities. While the item measuring the importance of reading and educational games is of substantive interest, given its high correspondence with the kinds of abilities assessed in these two cognitive tests, it is a less useful discriminator of views than the composite measure of broader views about the importance of activities. This is unsurprising as a very high proportion of parents rated educational games as very important. Nevertheless, even amongst those who said this was important the degree of importance they attached to it (very or quite) was correlated with ability scores.

Table 4.7 Average cognitive ability T-scores at age 34 months by parental attitudes at 22 months

Parental attitudes	Mean score	Bases
	Naming vocabulary**	Weighted		Unweighted
Importance of playing educational games/reading with child +
Very important	52.5	3198		3263
Quite important	47.5	598		574
No. of activities rated as important
Four or five	53.5	1958		2026
Three	51.7	933		957
Two	48.9	514		495
One or none	46.7	442		402
Satisfaction with child's activities
Very happy	53.3	736		794
Quite happy	52.7	1431		1478
Would like a slightly wider range	50.6	1048		1032
Would like a much wider range	49.1	594		538
	Picture similarities**
Importance of playing educational games/reading with child +
Very important	50.3	3197		3261
Quite important	47.1	598		574
No. of activities rated as important
Four or five	50.8	1957		2024
Three	49.5	934		958
Two	49.0	513		494
One or none	46.0	442		402
Satisfaction with child's activities
Very happy	51.3	736		794
Quite happy	49.8	1430		1476
Would like a slightly wider range	49.2	1048		1032
Would like a much wider range	48.3	593		538

+The sample sizes for the rest of categories are too small to report

**Differences significant at less than .001

4.5 Conclusion

This chapter has provided an overview of the relationship between cognitive ability and socio-demographic factors and children's activities. The results confirm much of what the existing literature suggests about the kinds of factors that influence cognitive outcomes. For example, children's individual characteristics such as their gender, birth weight and developmental history are all associated with their cognitive ability at age 34 months with boys, low birth weight babies and children who had experienced developmental difficulties in the first year of life all found to have below average ability scores. Children whose mothers were aged under 20 when they were born or who have low levels of educational achievement (two factors that are themselves strongly linked) also performed below the average. Household and area level indicators of disadvantage, such as low income, unemployment or area deprivation also showed negative associations with ability scores. The range of activity measures defined in the previous chapter were also found to be associated with cognitive ability; children who experience a wide range of activities and those whose parents consider this to be important had higher than average ability scores. Whether the fact that children from more advantaged backgrounds also experience more activities is the underlying explanation for this apparent association between activities and cognitive outcomes remains to be seen. The next chapter aims to untangle this.