ESSnet on Data Collection for Social Surveys Using Multi Modes (DCSS) ESSnet Workshop 2014 Valencia, Spain 23 24 January 2014 karen.blanke@destatis.de
Outline 1. Background and key aims 2. Challenges 3. Results and deliverables 4. Conclusion Slide 2
Background Initiated by Eurostat to support NSIs in: Developing web data collection tools Implementing and assessing the impacts of mixed mode data Labor Force Survey (LFS) chosen as a concrete example Duration of the project: autumn 2012 2014 Consortium: Partners (5): FI, NO, NL, UK, DE (coordinator) Support-group-members (3): SE, DK, IT External consultants Slide 3
Key aims TechnicalIy Developing web data collection tools Assessing the impacts of mixed mode data Strategically to stimulate early collaboration among countries harmonise and standardise national practices transfer knowledge, and to support less experienced countries Fullfilling (so far) Exchange of knowledge within the consortium Dissemination to the ESS Slide 4
Challenges (I) Inception phase To find and select partners NSIs don t want to join without getting reasonable paid Financial cut down and changing running time Project start was rather vague, due to signature of contract Involvement of other ESS members Communication & dissemination First year: Draw attention to the project Interested NSIs contacted via Eurostat/project-coordinator Dissemination via CROS-portal Final Workshop September 2014 Slide 5
Challenges (II) Deliverables of DCSS Documents via CROS-portal Findings are particular important for less experienced countries Experienced countries: Exchange of specific research is available on the same survey and topic Assessment 2013 Assessment is considered as necessary & professional No information about assessment before Wrong assessment, two examples: Classified as missing deliverables, but actually should be provided later (no relation to time schedule) Only based on public documents via CROS-portal However, via google assessment is public Slide 6
Query on Data Collection in Social Surveys Data collection: spring 2013 37 NSIs (European & overseas) 93 % response rate Topics: MM in LFS & other Social Surveys Web questionnaires design Software package & IT Future policy towards web quest. Slide 7
Selective results: LFS by modes PAPI BG EL HR PL RO CATI CH FI IS LU NO SE CAPI BE IE TR Additionally: 6 countries are using registers supplementary CAWI CATI DK CATI CAPI AT CY ES FR HU IT LV PT SI UK US CATI PAPI MK MT RS CAPI PAP DE CAWI CATI CAPI AU NL CATI CAPI PAPI CZ EE LT SK CAPI CATI PAP NZ Slide 8
Selective results: Census and modes PAPI EL HR CAPI CY TR PAP FR IE CAWI CAPI EE LV 2) CAWI PAPI LT 2) SK CAWI PAP CH 2) CZ CATI CAPI NL 2) PAPI OTHER BG CAWI CATI CAPI PL CAWI CATI PAPI ES CAWI PAPI PAP HU PAPI PAP CAWI DE MK LU IT MT RS PT RO UK 1)Additionally 8 countries (AT, BE, DK, FI SI, SE, IS NO) applying register as only data source. 2)Additionally using register Slide 9
E1: Policy towards the implementation of web data collection for social statistics Region Yes, within the next 5 years Yes, but not within the next 5 years Not (yet) decided No European countries 13 1 3 6 N = 24 Statistical Data Collection, 25-27 September 2013, Geneva Slide 10
Web data collection main challenges Complex in concepts Instructions are needed Long answering lists Error checks Respondent s expectation Easy and quickly to fill in http://www.thiossane.org/images/usability2.jpg Don t make me think (Krug 2006) Slide 11
Placement & design of explanations (DE) Text behind link Directly visible text Keyword and text behind link Combination of visible text and link Statistical Data Collection, 25-27 September 2013, Geneva Slide 12
Long lists & functionalities (DE) Slide 13
Don t know options (NL) 14
Don t know options (NL) 15
Smartphones (NO) Slide 16
Multimode data collection Mode effects: Review of current knowledge on mode effects How to analyse and control mode effects Research on mode effects for the main variables of the LFS Fieldwork and case management systems & It-environment Impacts on weighting and estimation Recommend methodological activities to further harmonisation of MM data collection in the European Statistical System Slide 17
Conclusion (methodological) Implementation of web questionnaires is sensitive: Very complex concepts in official Switch from interviewer administered to self-completion Web questionnaires become more and more important Internet penetration increases but web questionnaire won t be the only mode Mixed-mode is Multimode in most household surveys, but very divers Countries are used to mm but there is less known about the impact on data quality Mixed-mode is sensitive for case-management & IT Slide 18
Future steps Web questionnaires Guidelines on the design of web questionnaires in official social statistics Automatic coding of profession /economic sector Training seminars on designing web questionnaires Slide 19
Future steps Mixed-Mode data collection Further research on mm and data quality in in official social statistics (based on recommendations of DCSS (I)) The impact of mixed data sources: surveys & registers Software requirements and case-management (case-studies) Sustainability of ESSnet results: Launch projects applying recommendations Set up Center of Competence Support bilateral consultancy Slide 20
Further contact: Coordinator/WP II: Karen Blanke (karen.blanke@destatis.de) Federal Statistical Office of Germany WP III: Annemieke Luiten (a.luiten@cbs.nl) CBS Statistics Netherlands Eurostat: (http://www.cros-portal.eu)