Booth, Charlotte; Crawford, Claire; Rajah, Nasir; Silverwood, Richard J; Henderson, Morag; (2024) Examining sample representativeness and data quality in the linked Next Steps survey and Student Loans Company administrative data. (CLS Working Paper 2024/1). UCL Centre for Longitudinal Studies: London, UK.

Preview

Text CLS-Working-Papers-2024-1-Sample-representativeness-and-data-quality-in-the-linked-Next-Steps-survey-and-Student-Loans-Company-administrative-data.pdf - Published Version Download (560kB) | Preview

Abstract

Linked cohort and administrative data provide rich data resources with wide ranging research possibilities. Yet, it is important to understand the quality and representativeness of linked cohort and administrative data, in order to establish the reliability of estimates from these. At the Next Steps age 25 survey (2015/16) participants were asked for their consent to link their survey data to various administrative records. One of which was the Student Loans Company (SLC), who provided data for consenting participants on any student loan applications, payments, and repayments made across a fourteen-year period (from 2007-2021). We examined sample representativeness for the linked Next Steps-SLC data by comparing participant characteristics between those who did and did not consent to data linkage, for those who were and were not successfully linked, and relative to national population statistics where possible. Among age 25 respondents, certain groups were less likely to consent to data linkage, including those from ethnic minority groups and more disadvantaged backgrounds. In the limited instances where comparable national data was available, the linked sample was found to be reasonably representative of these wider populations. Data quality was examined by evaluating agreement between similar variables held across both sources, which on the whole, revealed a high level of agreement suggesting high data quality. Finally, a novel policy-relevant research question was investigated, to showcase some of the research possibilities of using this linked data. We conclude that future research should capitalise on this new linked data resource for investigating a range of outcomes among student loan borrowers but note that certain groups may be underrepresented in the linked data, due to differential linkage consent rates.

Download activity - last month

Export as JSONXMLCSV

Download activity - last 12 months

Export as JSONXMLCSV

Downloads by country - last 12 months

Export as JSONCSVXML

Archive Staff Only

View Item