Following up on the ESRA 2019 presentation“Using Big Data to Decrease Coverage – Error and Sample Bias in European Establishment Surveys” this year at the 9th ESRA 2021 conference that was held virtually on the 2nd of July 2021, our colleague Nikola Jovanovski presented the overview on the latest developments we made.
“Ways of reducing coverage and sampling error as part of the Total Survey Error Framework for Establishment Survey in Europe; recent developments” is the title of this year’s research paper which addresses 3 important research questions:
⚫ How can Big Data be used for increasing the telephone coverage in the Sampling Frame?
⚫ How can the Sampling frame go online, for a mixed mode sampling approach?
⚫ How can we ensure sample data accuracy and validity in an international and longitudinal study?
The paper outlines the enrichment methodology covering the whole frame for both online and offline mode by using the Full Sampling Method; It starts with the business register where multiple traditional sources are merged and blended into one, evaluate the variables available, enriching the unavailable data and where needed generating and validating email addresses in order to complete the full data set.
The benefits of using the proposed strategies compared to the traditional ones are detailed rich Sample Data that contains more data elements than any other individual sources.
⚫ Data accuracy is validated through multiple sources.
⚫ Combining online and offline mode in survey fieldwork which brings increased coverage as more Sample data can be reached via phone or email.
⚫ One centralized source containing comprehensive data for all EU 27 countries.
Looking back at the ESRA 2019 future action steps 3 out of 5 are fully accomplished and 2 are still pending. The ones that are completed are: Using Big Data to identify company size on businesses with unknown size, include a validation step in the enrichment process to lower data inaccuracy and last but not least source contact person data and email. From the pending ones, there is more work to be done on including an area typology stratification criteria in order to make a more representative sample to frame and design Sample of all countries from a centralized source.
Although it was not possible to physically meet with everyone attending the ESRA conference as previous years, we are glad to see everyone online, to learn about the rapid advances in survey research and how the developments have resulted in identifying best practices and new opportunities for methodological research.
Looking forward to the last batch of live presentations on the 23rd of July and hopefully at the next conference there will be a chance to meet the ESRA community live again.
