The impact of using web in The Danish LFS Labour Force Survey workshop 2018, Reykjavik Tine Cordes
Background – changes i 2016 Core individual survey: Approx. 19.000 responses quarterly out of a sample of about 34.000 individuals. Changes in data collection in 2016 Multimode data collection – from pure CATI to a combination of CATI and CAWI. New data supplier from 2016. Could not reach our target response rate of 60 % of the net-sample. Termination of contract with supplier with effect from 2017.
Background – changes in 2017 The data supplier from 2007-2015 resumed data collection from 2017. Increased response rate, reaching our target of 60 percent of net sample. Changed the share of CATI and CAWI to most CATI. Invitation letter send to digital mailbox instead of by ordinary post. Breaks in time series in 2016 and again in 2017.
Analysis: The impact of CAWI 1) Are CAWI-respondents different from CATI-respondents? Looking at essential background characteristic for CATI- and CAWI-respondents. 2) Does the survey mode affect the way people answer the questions about employment in the reference week? Benchmarking LFS-employees in CAWI and CATI with register-employees from our register based statistics on employees (RSE).
1) Are CAWI-respondents different from CATI-respondents?
1) Are CAWI-respondents different from CATI-respondents?
2) Bias in survey mode - method Statistics Denmark’s register based statistics on employees (RSE) contains information on all employees in Denmark. Objective benchmark to check employee status in LFS with RSE. Limit my population of analytical interest to employees according to the LFS and checking their correspondence with RSE. Use information from part 1 to make subgroups where CAWI- and CATI-respondents are much more alike.
2) Bias in survey mode
2) Bias in survey mode
2) Bias in survey mode – time aspect
2) Bias in survey mode – time aspect
2) Employed students – a looser attachment to the labour market
Conclusion The CAWI-set-up inspires to far more voluntariness in the participation to the survey. The largest part of these respondents answer within the first two days; they are older, more educated and to a larger extent women compared to CATI-respondents. Even when controlling for these differences the results from the different subgroups all point in the same direction: that CAWI gives more accuracy in measuring employee-status than CATI. The difference between CAWI- and CATI-employees in their correspondence with RSE was bigger for groups with a looser attachment to the labour market (students and young people). Being presented to a calendar in CAWI highlighting the reference week must play a significant role for these differences. Especially the calendar must be helpful the looser attachment the respondent has to the labour market and the longer time gap there is between reference week and completion of questionnaire.
Thank you for your attention! Tine Cordes tco@dst.dk