Harmonising the measurement of social origin, cognitive ability and educational attainment across the National Child Development Study (NCDS), the British Cohort Study (BCS70), the Longitudinal Study of Young People in England (LSYPE), and the Avon Study of Parents and Children (ALSPAC)

(2016) - B Betthäuser, M Bourne

Journal: SOCED Project Data Note
Link::
DOI::
Links::
Tags:: #paper #NCDS #BCS #LSYPE #ALSPAC #CognitiveAbility #Attainment
Cite Key:: [@Betthauser2016]

Abstract

This data note has been prepared for the research project on Social Origins, Cognitive Ability and Educational Attainment: A Birth Cohort and Life-Course Perspective. The document will present descriptive statistics comparing key variables across four different data sets: National Child Development Study (NCDS), British Cohort Study (BCS70), Longitudinal Study of Young People in England (LSYPE) and Avon Study of Parents and Children (ALSPAC) (University of London. Institute of Education. Centre for Longitudinal Studies, 2015, 2016, 2018; Boyd et al., 2013; Fraser et al., 2013). For detailed information on each of the datasets used please see Bourne and Betthaeuser (2016); Betthaeuser and Bourne (2016a, 2016b, 2016c); and Bourne (2016). Further information on the ASLPAC dataset, the ALSPAC data dictionary and detailed acknowledgements can be found in the Appendix.

Notes

“rincipal Components Analysis (PCA) was run on the main cognitive ability (or KS2 performance) variables. Scores from the first unrotated component extracted were saved for each case, thus providing a measure of each child’s cognitive ability otherwise termed ‘g’.” (Betthäuser and Bourne, 2016, p. 10)

“Eigenvalue, loading, percentage of variance across the measures used that is accounted for by the first principal component score, and a summary of the generated variable for each survey.” (Betthäuser and Bourne, 2016, p. 10)

“While Key Stage 2 scores and scores on tests of cognitive ability tend to be highly correlated, additional analyses carried out elsewhere have led us to conclude that the former cannot be considered a reliable proxy” (Betthäuser and Bourne, 2016, p. 11)

“Expectantly, cognitive ability scores tend to be normally distributed within each of the datasets. In our main analyses, we intend to group scores into cohort-specific cognitive ability quintiles in order to account for ‘Flynn effects’ (Flynn 1987) and any potential non-linear effects on educational attainment.” (Betthäuser and Bourne, 2016, p. 11)

“Table 8. Distribution of Educational Attainment at Age 20 (%, five category measure, up to and” (Betthäuser and Bourne, 2016, p. 12)

“Table 12 shows the rank correlations between the different social origin measures. Consistent for each dataset, the highest correlations exist between parental class and parental status, though these are always low enough to avoid posing problems of multicollinearity for our multivariate analyses” (Betthäuser and Bourne, 2016, p. 14)

“NCDS” (Betthäuser and Bourne, 2016, p. 14)

“Table 13. Cross-tabulation of Social Class by Educational Attainment (Row %) – NCDS” (Betthäuser and Bourne, 2016, p. 15)

“Table 17. Cross-tabulation of Social Status Quartiles by Educational Attainment (Row %) – NCDS” (Betthäuser and Bourne, 2016, p. 16)

“Table 21. Cross-tabulation of Parental Education by Educational Attainment (Row %) – NCDS” (Betthäuser and Bourne, 2016, p. 17)

“Average Centre Four Global Log Odds Ratios and Test of Significance in the Differences between Social Origin Measures” (Betthäuser and Bourne, 2016, p. 19)