Growing up in Scotland: change in early childhood and the impact of significant events

Footnotes

1. For further information about weighting in GUS see the data user guides on the GUS website www.growingupinscotland.org.uk

2. Families who could have experienced all events would have been job rich couple families where the mother was in good health at the time of the first interview.

3. Further detail is provided in Appendix A.

4. Early parent-child relationship is measured in year 1 using selected items from the Condon and Corkindale's Maternal Postnatal Attachment Scale (1998), see Appendix A for more information.

5. Such approaches can result in driver score at year 5 (the dependent variable) being highly correlated with the driver score at year 1 (one of the independent variables), which manifests itself in high values of R2 and makes it more difficult for other associations to be detected. However, this should be less of a problem given the relatively large time span between the two measurements (four years).

6. GUS records the number of natural parents living in the household at each sweep, and the term couple family in this chapter refers to families where both natural parents are living together with the GUS child. Table 3.1 excludes the few cases where the GUS child lived with neither natural parent at sweep 1.

7. The association between marital status and separation remains significant even when controlling for the duration of the relationship (analysis not shown).

8. This, as with the majority of questions in GUS, is asked to the main respondent (usually the mother). So there may be cases where the mother was not living with the child last year but now is, even though the children have not actually moved house. This is likely to be a very rare event, so we assume in our analysis that all changes in address apply to the GUS child.

9. Analysis of FACS data by the authors also shows a fall in the prevalence of house moves in the past 12 months by age of the youngest child in the household.

10. Every effort is taken to track people in the study and the relevant cross-sectional and longitudinal weights are used in all analysis in this report, however it is possible that the number of frequent movers is underestimated.

11. In 2004/05 when the GUS babies were born employed mothers were entitled to up to 52 weeks of maternity leave. Respondents on maternity leave are treated as economically active with their job status recorded as it was prior to starting maternity leave. As the first interview took place when the GUS child was 10 months old the higher prevalence of job loss between the first and second interviews may partly be due to some mothers having been on maternity leave at the time of the first interview from a (full-time) job to which she did not subsequently return.

12. Initial modelling of job loss including both sets of parents in the same model revealed significant interactions between family type and a number of the explanatory factors.

13. Other accommodation includes accommodation tied to employment, temporary accommodation and living rent-free, for example with the child's grand-parents.

14. The majority of lone parents were either out of employment through out the period (43 per cent) or out of employment at the beginning of the period and then entering employment, temporarily out of employment, or cycling in and out of employment during the period (27 per cent).

15. Families who could have experienced all events would have been job rich couple families where the mother was in good health at the time of the first interview.

16. Child Poverty Strategy, Scottish Government, http://www.scotland.gov.uk/Publications/2011/03/14094421/6

17. All of these items are used for the derivation of both the physical and the mental health scores but weights are applied to the items differently for the two components.