The international road to new forms of data collection

/ Author: Miriam van der Sangen
Early last month, Statistics Netherlands (CBS) hosted two UNECE workshops. More than 100 delegates from all parts of the world met in The Hague to discuss data collection and communication of statistics. Three key players during these events were Bilal Kurban of the Turkish Statistical Institute, Michael Levi of the US Bureau of Labour Statistics and Heini Salonen of the United Nations Economic Commission for Europe (UNECE).

Statistics in a changing world

The way in which statistical offices collect data is changing, and so is communication on statistical results with the outside world. These two topics took central place during the two UNECE workshops. From 3 to 5 October, the focus was on data collection; from 5 to 7 October, it was on statistical information dissemination and communication. In between, on 5 October, a joint event was held for the participants of both workshops, at which knowledge and ideas were exchanged.

Web-based data collection

Bilal Kurban, Head of the Central Data Collection Unit at the Turkish Statistical Institute, acted as chair of the first workshop. He is enthusiastic about the increase in electronic data acquisition at companies. ‘Collecting data by using internet surveys offers many advantages: convenience for the respondents, more up-to-date results as well as accuracy of data at relatively lower costs.’ For social statistics, more and more NSIs are using a combination of surveys on paper, by telephone and/or through the internet: the mixed mode approach. Kurban: ‘The main reason for this is that survey costs can be reduced while quality is maintained. The mixed mode approach does however require a more complex production process and proper monitoring, for example in order to make responses mutually comparable.’

Measuring the response burden

Kurban points out that nowadays, merely looking at time spent on the completion of questionnaires by companies and households does not suffice. ‘We must also take into account how they themselves experience the response burden. To measure this, we can make use of the paradata (i.e. data providing information about the context and the process, ed.) which are automatically generated during an online survey, including the time it takes to complete a questionnaire.’

Innovation in data collection

Kurban stresses that the problem of survey non-response among businesses and households is growing. Therefore, in order to maintain the reliability of statistical results, the process of data collection must be renewed continually. He cites initiatives by Statistics Finland, among others: ‘Statisticians there are trying to improve both survey response and quality of data by using visualisation techniques which monitor the data collection process. In addition, CBS is implementing innovative ways of tackling non-response, response burden and shrinking budgets. One way as demonstrated at the conference by CBS is Standard Business Reporting, which allows businesses to submit their data quickly and safely. Another topic was web scraping, i.e. extraction of data from websites.’ According to Kurban, new methods of data collection do place different demands on the staff: ‘They need to have both IT and statistical capacities. It is vital for statistical offices to help their staff develop skills in these areas. Online staff training courses such as those at CBS are a good way to acquire data science skills.’

Creativity and new insights

Michael Levi is in charge of the Publications programme at the US Bureau of Labour Statistics. He was closely involved in the organisation of the communication workshop. The workshop comprised various sessions, with topics including how to measure the impact of communication devices, interactive tools, open data, new apps and techniques to improve communication with the respondents. He is excited about the outcome: ‘What I find particularly interesting is how participants from different professional backgrounds can come to concrete solutions. Many of the sessions demonstrated remarkable insights and creativity.’ Levi mentions an impressive presentation held by Amaria Silvina Viazzi, Director of Dissemination at Argentina’s National Institute of Statistics and Censuses since last January. She gave an account of the crisis faced by the bureau and the invasive approach they have taken to get out of it. Levi furthermore commends the various innovations in digital publishing which participants shared with each other. ‘A great example is the new fully digital Eurostat publication on energy in the European Union.’

‘Many sessions demonstrated remarkable insights and creativity’

Experiment

UNECE organiser Heini Salonen is satisfied with the results from both workshops, especially of the joint event on 5 October. ‘This was when participants from both workshops met with each other. It was the first time for us to host a one-day event with people in both data collection and communication where we asked them to actively work together on group assignments. It was a kind of experiment aside from the usual listening to each other’s presentations. The workshops provided opportunities to discuss one another’s experiences and the lessons that can be learned from them.’ Among the topics of discussion was how to promote the use of mobile devices, and how this offers possibilities for questionnaire design and increase in response rates. ‘The workshops were an important catalyst for new ideas. At the end, one of the participants said he had gathered so many good ideas that he was not sure if he would have the time to present all of them to his colleagues.’