Chair: Marina Signore
Room: S2 Wawel
Time: 14:30 - 16:00
Date: 28 June
|Marina Signore/ Manuela Murgia |
|Title: <<< The MIMOD Project: an overview >>>
In response to a Eurostat call for a Grant on “Cooperation on Multi-Mode Data Collection”, a Consortium led by Istat (Italy) in partnership with CBS (Netherlands), SSB (Norway), STAT (Austria) and Destatis (Germany) was set up and the MIMOD – Mixed Mode Designs for Social Surveys was awarded the Grant. A network of supporting countries - INSEE (France), Czech Statistical Office (Czech Republic), Central Statistical Office of Poland (Poland), Statistic Finland (Finland) and Statistics Sweden (Sweden) - will provide inputs to the project. The MIMOD project aims at supporting NSIs in facing a range of challenges which are at the forefront of applied research when implementing multi-mode and multi-devices data collection. The MIMOD project covers the following main topics: I) mode organisation with the objective of determining the steps in a decision tree that supports the implementation of adaptive and responsive mixed-mode survey designs; II) mode bias/mode effect and its adjustment with the aim of providing general guidelines on methodologies to deal with mode effects in multi-mode designs; III) case management in mixed-mode data collection with the purpose of investigating the different systems in use in terms of technical components and organisational approaches used, as well as challenges in efficiency and quality ; IV) mixed-mode questionnaire designs in order to give best practice recommendations on approaches for developing questionnaires for mixed-mode surveys as well as on modes used in the contact and follow-up phases of data collection; and V) challenges for phone and tablets respondents in CAWI with the aim of investigating the use of mobile devices (smartphones, tablets) in ESS surveys, and of mobile device sensors (such as GPS, camera, microphone, accelerometers) to enrich ESS surveys.
The paper will describe the activities and the mid-term results achieved by the MIMOD project.
|Title: <<< Dealing with mode effects >>>
One specific workpackage of the European project MIMOD - MIxed MOde Designs for social surveys is devoted to the evaluation of methodologies for dealing with mode effects in surveys using multi-mode data collection. Recent literature widely discuss the problem of identifying and treating mode effect components, i.e. mode selection (resulting from errors of nonobservation), and mode measurement (resulting from observation errors - coverage, nonresponse, and measurement errors). The purpose of the workpackage is to look into ways to cope with mode effects (e.g. weighting, imputation, other methods), and to analyze differences in the final sample composition based on different modes across time, countries and survey types, providing practical, evidence based guidelines for the NSI's. In the first period, the activities will be focused on an updated literature review on methods for the assessment and the adjustment of mode effects in mixed-mode designs, particularly those currently used in the ESS, with a discussion of assumptions, advantages and disadvantages of the various approaches. This activity will be supported by the survey which will be conducted at the beginning of the projcet on a selected group of NSI's. The suitability of selected statistical approaches and methods for mode effect assessment and adjustment will be then evaluated based on applications on current multi-mode social surveys. Based on the previous activities, general guidelines about methods and approaches which can be adopted to deal with mode-effects in multi-mode designs will be provided. The paper will report some outcomes of the workpackage’s activities in the first half of 2018, including some results of the MIMOD query and main evidences from literature review. The overall strategy which will be adopted in order to assess mode effect estimation (and adjustment) in selected ESS surveys will be also delineated.
|Dag F Gravem|
|Title: <<< Adapting ESS survey questionnaires to mixed-mode data collection >>>
Many countries in the European Statistical System (ESS) are or have recently been in the process of transferring previous interview and paper surveys into web and mixed-mode surveys. A 2012 survey among 17 National statistical institutes (NSIs) showed that this process for many NSIs involved major adjustments of both wording, structure and placement of instructions. Shifting from interviewer administered surveys to self-administered in particular potentially can result in mode effects resulting in differences in estimates. No structured overview of such question changes or the effect they may have on data quality for key questions and surveys within the ESS exists. Work package 4 of the MIMOD (Mixed Mode Designs for Social Surveys) project aims to amend this by offering evidence-based recommendations for mixed-mode questionnaire adaption. This will be done by reviewing available documentation from NSIs, as well as through prestesting conducted specifically for the project. The paper will present the work package in more detail, including which surveys, questions and question types that are analysed and tested, and offer preliminary results from pretesting.
|Bart Bakker |
|Title: <<< ESS surveys and mobile device data collection >>>
Mobile device coverage has increased steadily and to data a substantial share of respondents attempts to complete online surveys on smartphones and tablets. Surveys in the European Statistical System have, however, rarely been made fit, let alone been optimized, for such devices. Apart from the completion of surveys on mobile devices, the devices also have another feature: They carry various sensors that facilitate automated measurements once respondents consent. In a number of surveys such as travel and health, such sensor measurements have already been tested and applied. In the paper, we discuss the application of mobile devices to ESS surveys: Are the surveys fit to be completed on mobile devices? Or can they be made fit with reasonable effort? And are sensor measurements useful and added value for the ESS surveys?