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

Thursday 16th July, 14:00 - 15:30 Room: HT-105

Web and mixed-mode data collection in National Statistics 2
Convenor	Mrs Karen Blanke (FSO Germany )
Coordinator 1	Mrs Annemieke Luiten (Statistics Netherlands)

Session Details

Many countries within the European Statistical System (ESS) are considering web-based data collection in a system of multiple mode data collection. Eurostat initiated the ESSnet project “Data Collection in Social Surveys Using Multiple Modes” with the purpose to support Member States in their development and implementation efforts. A consortium of five NSIs has done extensive research on the development of web-based data collection tools for NSIs, specifically the Labour Force Survey, and the impact of implementing multimode data collection.
Concerning the web questionnaire, much of the work was aimed at finding out how severe the challenges in switching to self-completion actually are and if/how interviewer-assistance can be adequately replaced in web questionnaires, in view of the complicated concepts measured. The second aim was how to design functionalities in web questionnaires like instructions, routing, edit checks, customised wording or the variety of question types.
Concerning mixed mode data collection, we focussed on three issues: the organisation of mixed mode data collection, mode effects and adjustment for mode and measurement effects. In ‘organisation’ we focus on mode strategies: which modes in which sequence, response rates and measures to heighten web response. Attention is also given to the important subject of case management systems: software systems that are able to support all modes and allow flexible transitions from one mode to another. Concerning mode effects, several studies were performed on mode effects in mixed mode LFS designs. The final research topic was estimation and adjustment: given that there are mode effects, can we adjust for them, and how should that be performed.
We propose either one session where the partners in this research project discuss the findings on the topics or alternatively, we could have a double session where a number of speakers are invited to present.

Paper Details

1. Mixed mode data collection in National Statistics. Findings from the ESSnet DCSS
Mrs Annemieke Luiten (Statistics Netherlands)

In the ESSnet project “Data Collection in Social Surveys Using Multiple Modes” extensive research is performed on web-based data collection tools for NSIs and the impact of implementing multimode data collection.
Concerning mixed mode data collection, we focussed on three issues: the organisation of mixed mode data collection (mode sequences, response rates, case management), mode effects and adjustment for mode and measurement effects. Several studies were performed on mode effects in mixed mode LFS designs. Finally, various methods to adjust for mode effects were evaluated. The presentation will summarize the findings.

2. Results and Data Analysis for the 2016 Fishing, Hunting, and Wildlife-Associated Recreation Survey’s Mixed-Mode, Pre-Screener Test
Mr Matthew Herbstritt (United States Census Bureau)

The U.S. Census Bureau must develop new methods to reach households for the 2016 National Survey of Fishing, Hunting, and Wildlife-Associated Recreation (FHWAR) while keeping the address-based survey within budget. The15,000 household pre-screener test was implemented in two vastly differently states. The test included an embedded experiment between a paper only questionnaire, a web questionnaire, and a paper/web option. The pre-screener test results were positive with response rates between 21 and 32%. The test and research findings indicate the pre-screener will likely produce higher contact rates thus saving time, money and resources.

3. Data Collection in Social Surveys in Latvia Using Multiple Modes Data collection
Mrs Maranda Behmane (CSB of Latvia)

One of the most serious problems in the area of social statistics in Latvia is the decrease of response in household surveys that as a result essentially influences the quality of the obtained information. One of the solutions – to use wider in the household surveys the possibility to fill in questionnaires in the internet. In Latvia the internet is used by 72% of households. It means that a wider use of CAWI questionnaires is one of the most important tasks in the nearest future.
In 2014 CAWI was used in the European Public Health Survey.