SSI

ABSTRACT:

Project SSI, Smart Survey Implementation, is a consortium of the core countries of ESSnet Smart Surveys. It consolidates ESSnet findings to a mature baseline set of smart surveys including PDCA-cycles, creates shared smart microservices, tests microservices in six countries based on four existing solutions, optimizes push-to-smart recruitment and motivation strategies, implements machine learning retrainig strategies, and advances both the conceptual framework and range of smart features in methodology, IT architecture, logistics and legal levels.

AIM (WHAT)

The SSI project seeks out to answer an EU call for the creation of a data collection process that provides reliable and comparable statistical information to European institutions, specifically in the form of Trusted Smart Surveys that: (i) involve citizens as active contributors; (ii) acquire, process, and combine data collected from smart devices and other appliances; and (iii) ensure trustworthiness by guarantying strong privacy safeguards. Within this context, the four concrete objectives of SSI are: (i) consolidation of what has been learned into a robust and flexible baseline smart survey design; (ii) preparation of the baseline for changes and extensions in time; (iii) exploration and inclusion of new smart features; (iv) adoption of a respondent perspective. Smart surveys are promising in cases when questions are burdensome or cognitively demanding, when they are poor proxies of concepts of interest, or when they require expert knowledge or detailed recall.

METHODOLOGY (HOW)

The methodological approach of SSI involves the following components:

• A cross-national survey on smart survey perceptions in three countries (Italy, The Netherlands, and Slovenia) with the aim of identifying (in)attractive features of smart surveys that might foster or hinder respondents’ participation;
• Maturity criteria that determine the decisions made in the development cycles of smart features;
• Expansion/elaboration of the GSBPM (Generic Statistical Business Process Model) based on a taxonomy of smart surveys. Evaluations/interviews with, among others, case management staff and data scientists will also be used to inform the extension and elaboration of GSBPM;
• Small-scale focus groups and evaluation sessions with target populations to identify attractive smart survey features that should be made salient and objections that should be addressed in recruitment and motivation tactics;
• Small-scale usability testing focusing on the respondents’ interactions with the smart features that have been developed and implemented;
• Development, implementation, and evaluation of full open source end-to-end shared smart services/libraries;
• Development of retraining (active/online learning) strategies for machine learning routines to account for time change with the help of Privacy Ensuring Techniques (PET);
• Field tests in different countries to evaluate recruitment and motivation strategies, as well as method effects between smart and non-smart implementations;
• Penetration tests and other audits to ensure data security.

IMPACT (WHY)

The SSI project will demonstrate maturity of smart survey solutions and will create strategies to maintain such solutions and to expand them with new features. The project will assist the ESS (European Statistical System) by supporting national statistical institutes in embedding shared smart microservices and libraries in their production of ESS surveys. In order to maximise their implementation and impact, the SSI outputs will be agnostic to a specific application and will be tested and evaluated for multiple applications. In addition to statistical institutes, scholars/researchers in the area of data collection innovation will benefit from the project results.