Smart surveys are surveys that make use of sensors and machine intelligence to reduce respondent burden and increase data quality. Smart surveys have been tests as a way to improve diary surveys in official statistics, where data are collected on topics such as travel, time use and household expenditures. There are often inherent differences both in measurement and representation between smart surveys and traditional diaries, which makes it difficult to integrate both data sources in producing statistics over time, or within a mixed- or multi-source context. This paper distinguishes two different approaches to integration: the mixed-mode approach, which prioritizes outcome alignment and minimizes measurement differences for straightforward data merging, and the multisource approach, which maintains inherent mode differences and integrates data at the modeling stage, allowing exploitation of the strengths of each source. Using travel surveys as an illustrative example, we explore the benefits and drawbacks of each approach, and propose a decision framework to guide researchers in selecting the appropriate integration strategy.
翻译:暂无翻译