Scottish Government

Chapter 4 Effects of multiple childcare use on child outcomes

4.1 Key findings

• For the birth cohort, analysis was undertaken to explore the independent association between various features of childcare arrangements experienced at age 10 months on cognitive development at age 34 months whilst controlling for key socio-economic characteristics which are known to influence cognitive ability in the early years.
• Of the various childcare characteristics at age 10 months considered, only weekly duration of non-parental care had any statistically significant association with the child's cognitive ability at age 34 months after controlling for key family socio-economic and demographic factors; non-parental care of between 17 and 40 hours per week was found to have a significant positive impact on a child's knowledge of vocabulary specifically amongst girls.
• The characteristics of childcare arrangements in the first year of life which could be considered to describe 'childcare fragmentation' - exposure to multiple providers, a greater mix of provision, and less time with any single provider - do not impact positively or negatively on child cognitive development at age 34 months.
• For the child cohort, the association between childcare features at age 34 months and behavioural development at age 58 months was explored.
• After controlling for key family characteristics such as parental education levels and parental employment, experiencing 40 hours or more of care per week at age 34 months was detrimental to children's behavioural outcomes as they approached their fifth birthday. Further analysis suggested this relationship was significant particularly for girls and for children whose mothers were under 25 at the child's birth. No other childcare features were significant.

4.2 Cognitive development

Cognitive ability in the birth cohort was measured at age 34 months by two assessments: the naming vocabulary and picture similarities subtests of the British Ability Scales ( BAS). Each subtest is part of a cognitive assessment battery designed for children aged between 3 and 17 years (Elliott, 1983). The assessments are individually administered.

Naming vocabulary requires the child to name a series of pictures of everyday items and assesses the expressive language ability of children. In the picture similarities assessments children are shown a row of four pictures on a page. They are asked to place a free-standing card with a fifth picture underneath the picture with which the card shares a similar element or concept. Picture similarities measures a child's problem solving ability. There are 36 items in total in the naming vocabulary assessment and 33 items in the picture similarities, however the number of items asked to each child is dependent on their performance. For example, one of the criteria for terminating the naming vocabulary assessment is if five successive items are answered incorrectly.

The analysis used normative BAS scores, derived from the standard BAS tables and defined with reference to the standardisation samples used in developing the assessments. An increase in either scale denotes an increase in cognitive ability.

Initial analysis explored the difference in average scores across key sub-groups defined according to various childcare arrangements and situations that have been defined in sections 2 and 3 above. Subsequent analysis then considered the independent effect of various childcare factors at age 10 months on cognitive outcomes at age 34 months whilst controlling for key socio-economic characteristics which are known to influence cognitive ability in the early years.

4.2.1 Mean scores by use of childcare

Initial analysis examined the differences in average ability scores on both scales according to whether or not the child experienced any form or duration of non-parental care at each wave of data collection. The results are detailed in Table 4.1

Table 4.1 Average cognitive ability scores by use of childcare at each wave of data collected

***Differences significant at less than .001

The data indicates that those children who had experienced regular non-parental care at any sweep had, on average, higher ability scores on both measures at age 34 months than those children who were 'home-reared', although the average differences are small.

To investigate the potential effect of long-term experience of non-parental care, ability scores were compared according to the pattern of childcare reported over the period
0-3 years (Figure 4 A). The results indicate that those children who were in non-parental care consistently over the three year period showed the highest average scores on both scales and those children who had not been placed in any childcare arrangements during that time showed the lowest average scores. Thus, on the basis of the measures used and this analysis, childcare use is supportive, on average, of children's cognitive development. More detailed analysis is carried out later in the chapter (section 4.2.5) which controls for other factors such as family characteristics which are known to be correlated with cognitive outcomes.

Figure 4 A Average naming vocabulary ability score at 34 months by pattern of 'any' childcare use between 0-3 years

4.2.2 Mean scores by number of childcare providers

We now turn to whether there is any relationship between the use of multiple childcare providers and children's cognitive development. Average ability scores were compared for children who received their care from different numbers of providers at age 0-1.

Table 4.2 Average cognitive ability scores at age 34 months by number of childcare providers in use at age 10 months

***Differences significant at less than .001

Whilst in the separate tests, the difference in average score between those with no childcare and those with any childcare is significant, there is no significant difference in cognitive ability at age 34 months between children who are cared for by different numbers of childcare providers in their first year. This data would suggest, therefore, that experience of multiple childcare provision in the first year of life has no, or little, detectable impact on child cognitive development by age 34 months. Although not shown in the table, further analysis confirms the results are similar according to which age point childcare provision is assessed. That is, there are no statistically significant differences in average cognitive ability scores between children who receive non-parental care from a single or multiple providers at either age 22 or 34 months.

Again, to explore any potential effect of long-term experience of multiple childcare provision, more detailed patterns of childcare use were mapped out for the period 0-3 years. This identified at which points children were looked after by two or more providers as detailed in Table 2.4. No significant differences were detected in mean scores on the picture similarities assessment according to patterns of multiple childcare use in the child's first three years (Table 4.3). However, differences in naming vocabulary scores were significant; those children who had experienced non-parental care by two or more providers at all ages showed the highest mean score. They were closely followed by children who were cared for by two or more providers at age 34 months. Children who experienced intermittent multiple care, at age 10 months and again at age 34 months, had the lowest average scores. From this we can conclude that the use of multiple childcare providers does not have an adverse impact on children's cognitive development, and in some circumstances, it has marginally beneficial effects.

Table 4.3 Average cognitive ability scores at age 34 months by patterns of use of two or more childcare providers

**Differences significant at less than .01

4.2.3 Mean scores by mix of childcare provision

We may also ask what are the consequences for children's cognitive development of care by different types of providers differentiated by whether formal or informal or group or individual care. Mean scores on both assessments were compared according to the informal/formal characteristics of provision in place for children at age 10 months.

The results in Table 4.4 show that those children who experienced a mix of formal and informal childcare at age 10 months, on average, scored slightly higher on each assessment than did children who experienced only formal or informal care. However, the difference in naming vocabulary scores between those who experienced only formal care and those in mixed care is extremely small. Again, these results should be treated with caution as the differences are small and the analysis does not control for family characteristics. Previous analysis has shown that there are key differences in the family characteristics of those who use different types of care. For example, lone parents, younger mothers and those on low incomes are considerably more likely to use informal care alone, thus the children in this group will be distinct in a number of important background factors likely to affect their cognitive development at this stage.

Table 4.4 Average cognitive ability scores at age 34 months by formal/informal characteristics of childcare at age 10 months

***Differences significant at less than .001

4.2.4 Mean scores by total hours of childcare per week

Data in section 3.2 above illustrated that children who were cared for by more than one childcare provider were more likely to spend longer periods in non-parental care per week. Average ability scores were compared according to the duration of weekly childcare experienced at age 10 months.

The results (shown in Figure 4 B) suggest that, to a degree, a greater duration of weekly non-parental care at age 10 months leads to greater cognitive ability at age 34 months; those children who spent between 17 and 40 hours in non-parental care at 10 months showed the highest average scores on both assessments. However, experiencing more than 40 hours of childcare appeared to be detrimental for cognitive development as those children had lower average scores than their peers in the 17 to 40 group.

This analysis does not particularly account for the use of multiple childcare providers as the majority of children at age 10 months were looked after by a single provider. However, section 3.2 noted that, for children with multiple childcare providers, duration of care in any single care environment was lower than for children with just a single provider. To explore, in simple terms, whether having more or less time in a single care environment appeared to have any effect on cognitive development average ability scores were compared according to the number of hours per week the child spent with each childcare provider (Table 4.5).

Figure 4 B Mean cognitive ability scores at 34 months by total weekly duration of childcare at 10 months

Table 4.5 Average cognitive ability scores at age 34 months by average number of hours in the care of each childcare provider at age 10 months

***Differences significant at less than .001

There were no statistically significant differences in scores on the picture similarities assessment according to the average weekly time spent with each childcare provider. However, on the naming vocabulary assessment children with higher average durations per provider showed higher average scores which may suggest that multiple childcare provision where only a small amount of time is spent with a greater number of childcare providers, is less beneficial to early cognitive development than is a longer time with a single or fewer providers.

4.2.5 Examining the independent effect of childcare fragmentation on early cognitive development

Previous research on child cognitive development has identified a number of key individual and family characteristics which impact on children's early attainment and could therefore affect their cognitive scores at age 34 months. As well as the child's gender, these also include poverty and socio-economic factors, parental education, family circumstances and marital breakdown and ethnicity. Previous analysis of GUS data, as noted earlier, has shown that patterns in use of childcare, the type of provision used, and, to a lesser extent, the number of providers used, also vary significantly according to many of these characteristics. Thus, in the outcome analysis undertaken thus far it is difficult to establish whether it is features of childcare use which are impacting on child cognitive development or the particular socio-economic and demographic characteristics of the families and children who are most likely to experience multiple and varied childcare provision.

Therefore, to more clearly identify any potential impact of the various features of childcare, linear regression analysis was undertaken which allows the individual effect of childcare variables on cognitive scores to be considered whilst controlling for key family characteristics. Explanatory factors considered in combination in the analysis included: the child's gender, the highest parental educational qualification in the household, the highest parental socio-economic classification in the household, the level of household income, parental employment status, lone parent status and ethnicity. These were considered alongside those childcare characteristics at age 10 months shown to be significant in the previous analysis including: number of childcare providers, informal/formal mix of childcare provision, total weekly duration of non-parental care (hours) and average number of weekly hours per provider.

The regression analysis showed that of the various childcare characteristics at age 10 months considered, only weekly duration of non-parental care had any statistically significant association with the child's cognitive ability at age 34 months after controlling for key family socio-economic and demographic factors ⁶. Non-parental care of between 17 and 40 hours per week was found to have a significant positive impact on a child's knowledge of vocabulary. Indeed, further sub-group analysis suggested that this effect was restricted to girls only ⁷. This suggests that, on the whole, the characteristics of childcare arrangements in the first year of life which we have considered to describe 'childcare fragmentation' - exposure to multiple providers, a greater mix of provision, and less time with any single provider - do not impact positively or negatively on child cognitive development at age 34 months.

4.3 Social, emotional and behavioural development

Social, emotional and behavioural development was measured in the child cohort at age 58 months (sweep 3) by administration of the Strengths and Difficulties Questionnaire (Goodman, 1997).

The Strengths and Difficulties Questionnaire ( SDQ) is a brief behavioural screening questionnaire designed for use with 3-16 year olds. The scale includes 25 questions which are used to measure five aspects of the child's development: emotional symptoms, conduct problems, hyperactivity/inattention, peer relationship problems and pro-social behaviour. A score is calculated for each domain, as well as an overall 'difficulties' score which is generated by summing the scores from all the scales except pro-social. For all scales, except pro-social where the reverse is true, a higher score indicates greater evidence of difficulties. The data was obtained via parental report, normally the mother, in the computer assisted self-completion module of the sweep 3 interview.

The analytical approach used was similar to that applied to the cognitive data. Initially, differences in average scores on each of the five sub-scales and on the overall difficulties score were explored across key sub-groups defined according to various childcare arrangements and situations. Further analysis then considered the independent effect of various childcare factors at age 34 months on SDQ scores at age 58 months whilst controlling for key socio-economic characteristics which are known to influence social development in the early years.

4.3.1 Mean scores by use of childcare

Initial analysis explored the differences in scores on all scales at age 58 months according to whether or not the child experienced any form or duration of non-parental care at age 34 months. The results are detailed in Table 4.6.

On this very simple examination, experience of childcare at age 34 months appears to have no bearing on developmental scores at age 58 months. The average scores returned on all of the scales amongst children in each group are almost identical.

Table 4.6 Average SDQ composite and difficulty scores at age 58 months by use of childcare at each wave of data collection

*Differences significant at less than .05

SDQ scores were also compared according to the pattern of childcare reported over the period 3-5 years (sweeps 1 to 3). The results indicate no statistically significant differences in behaviour scores between children who had experienced non-parental care at any single sweep and those who had experienced non-parental care at two or three sweeps.

4.3.2 Mean scores by number of childcare providers

We now consider whether there is any relationship between the use of multiple childcare providers and children's behavioural development. Average SDQ scores at 58 months were compared for children who received their care from different numbers of providers at ages 34, 46 and 58 months (Table 4.7).

Table 4.7 Average SDQ composite and difficulty scores by number of childcare providers being used at each sweep

No statistically significant differences were observed in the scores of children who experienced non-parental care from different numbers of providers at any age. This suggests, therefore, that experience of multiple childcare provision at age 2-3, 3-4 or 4-5 years has no, or little, detectable impact on child social, emotional or behavioural development by age 58 months. Looking at data across the three time points, the cumulative experience of multiple provision did not appear to be important either. There were no significant differences in average scores between those children who had never experienced care from three or more providers in the period considered, those who had experienced it only at one sweep, and those who had experienced it at more than one sweep.

4.3.3 Mean scores by mix of childcare provision

As with cognitive development, it is worth considering the possible consequences for children's social development of care by different types of providers differentiated by whether formal or informal in nature. Mean scores on the various scales were compared according to the informal/formal characteristics of provision in place for children at age 34 months (Table 4.8). No statistically significant differences were observed.

Table 4.8 Average SDQ scores at age 58 months by formal/informal characteristics of childcare at age 34 months

4.3.4 Mean scores by total hours of childcare per week

Average SDQ scores were compared according to the duration of weekly childcare experienced at age 34 months (Table 4.9). Differences in scores on the emotional symptoms, conduct problems, hyperactivity and total difficulties scales were statistically significant. The data shows a similar pattern to that observed in the cognitive data above where those children who experience extremely long durations of weekly care are distinct. In this instance, children who had 40 or more hours of non-parental care at age 34 months had higher average scores on all scales suggesting they exhibited more problematic behaviour at age 58 months than did children with shorter durations of care or who had no childcare at all. Furthermore, as with the cognitive data, durations of
non-parental care of between 17 and 40 hours per week appear to be most beneficial for social development with children in this group having the lowest average difficulty scores.

Table 4.9 Average SDQ scores at age 58 months by weekly hours in childcare at age 34 months

*Differences significant at less than .05
**Differences significant at less than .01

4.3.5 Examining the association of multiple childcare provision on social and behavioural outcomes

The previous analysis has shown that children's experience of non-parental care via multiple providers at age 34 months bears little positive or negative impact on their social and behavioural outcomes at age 58 months. Indeed, of the various childcare characteristics considered at age 34 months, only weekly duration of care was statistically significant; children who experienced more than 40 hours of non-parental care per week scored higher on a number of individual SDQ scales as well as the overall difficulties scale. This suggests that these children exhibit more problematic behaviour at age 58 months than their peers who experienced shorter periods of weekly non-parental care.

As discussed in relation to cognitive outcomes, previous research identifies a range of key individual and family characteristics which impact on children's early development and which could affect their behaviour at age 58 months. Further analysis was undertaken to test whether the relationship between long durations of care at age 34 months and more problematic behaviour at 58 months persisted when these characteristics are controlled for. The analysis showed that even after controlling for key family characteristics such as parental education levels and parental employment, experiencing 40 hours or more of care per week at age 34 months was detrimental to children's behavioural outcomes as they approached their fifth birthday ⁸.

It may be possible that this relationship between childcare duration and child outcomes differs for different groups. To test this, the full specification model was run separately for different groups of children based on the child's gender, parental education, family type, maternal age at the child's birth and level of household income. The analysis found that the relationship between duration of weekly childcare and more negative behavioural outcomes is significant particularly for girls ⁹ and for children whose mothers were under 25 at the child's birth ¹⁰. The former finding is interesting in particular because, in general and across numerous research studies, girls tend to score lower on the SDQ than do boys, suggesting they exhibit fewer difficulties. Further examination of these associations is beyond the scope of this paper, but it may be assumed that differences in the precise characteristics of the care experienced by the different groups of children may provide some explanation, particularly for those children with younger mothers whose care is usually informal in nature.