ESRA logo
Tuesday 14th July      Wednesday 15th July      Thursday 16th July      Friday 17th July     

Thursday 16th July, 14:00 - 15:30 Room: HT-103

Methodological issues of using administrative data to improve the quality of survey data 1

Convenor Dr Emanuela Sala (Dipartimento di Sociologia e Ricerca Sociale, Università di MIlano Bicocca )
Coordinator 1Dr Jonathan Burton (ISER, University of Essex)
Coordinator 2Dr Gundi Knies (ISER, University of Essex)

Session Details

There is a body of research on eliciting respondents' consent to link their survey data to their administrative records. However, the process of linking survey and administrative data is complex and the issues that survey practitioners need to tackle go beyond the asking for respondents' consent. Furthermore, using that linked data also provides potential for further methodological research. The aim of this session is to foster discussion on i) methodological issues that concern the data linkage process, ii) the research potential of the linked data,
iii) use of administrative data to improve the quality of survey data

We welcome papers on the following topics:

1. Analysis of mis-reporting
2. Measurement 'bias' using linked data
3. Impact on research findings compared to using survey data only
4. Validation studies
5. Methodological lessons from linkage process
6. Implications of consent bias

Relevant papers on other aspects of data linkage will also be considered.

Paper Details

1. Interviewers' influence on bias in reported income
Dr Manfred Antoni (Institute for Employment Research (IAB))
Ms Basha Vicari (Institute for Employment Research (IAB))
Mr Daniel Bela (LIfBi / NEPS)

Questions on sensitive topics like income often produce considerable item-nonresponse or measurement error. While several studies examine item-nonresponse little is known about misreporting. Misreporting might be explained by social desirability bias, which may lead to overreporting or underreporting of characteristics.
We use interviewer and respondent characteristics to estimate their impact on the extent of deviation between reported and recorded incomes. To do so we combine data from the German National Educational Panel Study’s adult cohort with administrative data from the Federal Employment Agency. These longitudinal data stem from mandatory notifications of employers to the social security system.

2. Using Information from Credit Records to Improve Survey Data
Mr Brian Bucks (Consumer Financial Protection Bureau)
Mr Mick Couper (University of Michigan)

We consider how credit records may be used to gauge and enhance survey data quality. The analysis focuses on a recent survey about debt collection which used de-identified credit records as a sampling frame, including to oversample consumers considered more likely to have debt collection experiences. The extensive information in the credit-bureau data for both respondents and nonrespondents provides a strong basis for investigating nonresponse bias. Comparing self-reported demographic data for respondents with auxiliary demographic information in the credit-bureau data sheds light on the reliability of such auxiliary data. Similarly, we compare self-reported financial data.

3. Recall Error in the Year of Retirement
Ms Julie Korbmacher (Munich Center for the Economics of Aging )

This project analyzes recall error in the year of retirement by comparing self-reports of respondents of the Survey of Health, Ageing and Retirement in Europe (SHARE) with administrative records of the same person provided by the German Pension Fund. A comparison of the two data sources show that 36.5 % of the respondents misreported the year they retired. Based on research from cognitive psychology, different determinants to explain the error are identified. The results show that cognitive abilities and characteristics of the event (as its salience or the time passed since the retirement) are correlated with the recall error.

4. Recall error in life course data – a comparison of survey data and administrative data
Dr Dina Frommert (DRV Bund)

The paper examines recall error in life course data by comparing survey data and administrative data month by month. A multilevel approach is used to determine the relative importance of factors relating to the errors, such as length of the recall period or length of the episode, and characteristics of the respondents, such as gender, education level, age, region and structural complexity of the life course.