Mostafa, T; Narayanan, M; Pongiglione, B; Dodgeon, B; Goodman, A; Silverwood, R; Ploubidis, G; (2020) Improving the plausibility of the missing at random assumption in the 1958 British birth cohort: A pragramatic data driven approach. (UCL Centre for Longitudinal Studies Working Paper Series 2020/6). UCL Centre for Longitudinal Studies: London.

Preview

Text CLS WP 2020_6 Improving the plausibility of the missing at random assumption NCDS.pdf - Published Version Download (4MB) | Preview

Abstract

Making the Missing At Random (MAR) assumption more plausible has implications for missing data analysis. We capitalise on the rich data of the National Child Development Study (NCDS - 1958 British birth cohort) and implement a systematic data-driven approach to identify predictors of non-response from the 11 sweeps (birth to age 55) of the NCDS (n = 17,415). We employed parametric regressions and the Least Absolute Shrinkage and Selection Operator for variable selection. Disadvantaged socio-economic background in childhood, worse mental health and lower cognitive ability in early life, and lack of civic and social participation in adulthood were consistently associated with non-response. Using this information, we were able to restore the composition of the NCDS samples at age 50 and age 55 to be representative of the study’s target population, using external benchmarks, and according to a number of characteristics captured within the original birth sample. We have shown that capitalising on the richness of NCDS allowed us to identify predictors of non-response that improve the plausibility of the MAR assumption. These variables can be straightforwardly used in analyses with principled methods to reduce bias due to missing data and have the strong potential to restore sample representativeness.

Download activity - last month

Export as JSONXMLCSV

Download activity - last 12 months

Export as JSONCSVXML

Downloads by country - last 12 months

Export as CSVXMLJSON

Archive Staff Only

View Item