enEnglish
CROS

MINTSE-NET (Minimize Total Survey Error Network)

(A) Presenter

Marc Plate, Statistics Austria

(B) Category

Early-stage idea

(C) Partners needed

Survey designers, IT experts, data analysts from Eurostat, NSIs, Universities 

(1) Objective

Provide tools to assess and minimize the Total Survey Error for all surveys in Europe and enhance survey capacity in the public domain

(2) Need/rationale 

public domain producers of survey data face an imbalance of increasing user demands and capacity. Each found their own solutions to design surveys and control their various sources of sampling and non-sampling errors. Joining institutional forces is imperative in view of resource constraints and the potentially massive gains from collaboration. However at present too little exchange and sharing is ongoing to meet the requirements of the data revolution, which has most recently been recognized by the UN-Secretary General.

(3) Points of departure 

Main points of departure may be seen in the approaches followed by the landscape of comparative surveys of the European Statistical System on one side (including e.g. LFS, SILC, AES, HIS, HBS, TUS, ICT etc...) and comparative surveys which are rooted in the academic or political domain (including e.g. European Social Survey, HBSC, SHARE, EQLS, MIDIS etc…). To these more conventional undertakings needs to be added a recent development of emerging online panels such as the LISS Panel run by Centerdata/ Tilburg University which have great merit for testing and advancements in survey methodology. Each survey and each domain have established some mechanisms of coordination and standardisation of survey design. What is sought is a more cross-cutting pooling of expertise, tools, standards and even staff to the benefit of all surveys in the public domain.

(4) A broad spectrum of methods 

i) thorough stocktaking and potentially empirical surveying of the nature of the different agencies involved, including their funding structures, strengths and weaknesses
ii) presentation of a common framework within the Total Survey Error paradigm
iii) production of an inventory of existing comparative surveys with indicative data on their errors
iv) recommendations to minimize the Total Survey Error, including the optimal composition of the survey landscape in the public interst
v) creation of comparative online panels following the design of the Dutch LISS to enable continuous methodological experimentation and study of design related survey errors 
vi) auxiliary measures aimed at enhanced training, staff mobility and common development teams or even common (web or CATI) data collection infrastructures

(5) Resources mobilized: 

The European Statistical System has excellent user relations, access to powerful infrastructures, vast technical expertise, quality assurance and census data and sampling frames. Its institutions do however often lack “research and development” to enforce their modernisation attempts. The flexibility, innovative spirit, the embracing of experiments and academic orientation which is often found in comparative survey research outside statistical offices could perfectly complement the strengths of official statistics for mutual interest.

(6) Expected results

Within a short period of time an efficient infrastructure is to be created which is open to all statistical producers in the public domain. It will act as a methodological service centre and broker interests and expertise which are currently fragmented across diverse institutional settings.