Survey Combining Multi-mode Data Collection and Big Data Extracts - - PowerPoint PPT Presentation

Towards an integrated Consumer Expenditure Survey

Combining Multi-mode Data Collection and Big Data Extracts

Gustav Haraldsen, Sverre Amdam, Li-Chun Zhang

2

CES 2012

• The diary is the most burdensome and costly part of the design.
• Response burden affect both quality and actual cost in a negative way
• The diary information cover a rather small part of the consumption

(1) E-receipts (2) Web Diary & Questionnaire

2017 Mixed Mode Survey Design

Inputs:

2017 Survey Quality Challenges

• Challenges

– Recruitment 1

• Peoples willingness to

disclose card transactions

10 20 30 40 50 60 Yes No Not sure

Willingness to disclose card transactions

• Recruitment 2

– Limited share of shops have yet adopted the e-receipt technology (e.g. grocery retail chains) – Loyalty programs are an alternative

Register Assisted Survey

Unit = Transaction Variable = Card Unit = Person/HH Variable = Characteristics Sampling Editing Sampling Integration Audit sampling

Survey Assisted Register Data

Extended TSE Model

Groves et al. revisited Transformation and Integration

Measurement Representation

Strategic Perspective

Ethical  Legal Technical  Competence GSBPM  Reorganization

Thank you

Gustav Haraldsen: GHa@ssb.no Sverre Amdam: Sad@ssb.no Li-Chun Zhang: LCZ@ssb.no Extended TSE Model Monetary Cost Perceived Response burden Relevance