Assessing dissimilarity of employment history information from survey and administrative data using sequence analysis techniques

Abstract

Life course data is frequently gathered either using retrospective surveys or linking records with administrative data. Yet, each strategy has specific advantages and disadvantages. We study the consistency between both types of data sources and reasons for mismatch using the linked data set SHARE-RV, which combines retrospective life history data from the Survey of Health, Ageing and Retirement in Europe (SHARE) with respondents’ administrative data from German pension insurance records (N = 1679). Utilizing sequence analysis techniques with Hamming distance, Optimal Matching and OMspell as matching algorithms, we examine mismatches between survey and administrative data covering detailed, 30-year employment histories, and analyze how inconsistencies are associated with life-course characteristics, demographic and socio-economic factors. Our results show that life-course complexity and spells of atypical employment are associated with more mismatches. Furthermore, gender differences are pronounced and appear to be sensitive to the applied matching algorithm.