Method effects in smart surveys
Conference
65th ISI World Statistics Congress 2025
Format: IPS Abstract - WSC 2025
Keywords: big-data, data-collection, official-statistics, surveys
Session: IPS 863 - Smart Survey Methodology
Thursday 9 October 10:50 a.m. - 12:30 p.m. (Europe/Amsterdam)
Abstract
Traditional data collection in surveys typically result in a rectangular dataset with rows (people) and columns (variables). Over the last few years there is an increasing interest in smart surveys, in which survey data are enriched and integrated with various sorts of Big Data, such as pictures, movement data, or location data in the data collection process Smart surveys therefore result in a much more complex data structure. This also has consequences for settings in which a new method for data collection via smart surveys has to be integrated with traditional surveys. In this talk we discuss two possible approaches: First a calibration approach, in which one of the methods (the new or the old), is restructured in such a way that data can be compared, calibrated, and combined. Second, we discuss a data integration approach, in which the different data sources are processed differently, and only outcome statistics are integrated at the very end. We illustrate both approaches with data from the Smart Survey Implementation project, that focused on producing official statistics using a smart survey approach for household budget and time use. We discuss advantages and disadvantages of both approaches and conclude with a discussion of implications for data collection in official statistics.