Smartphones: collecting better social data

Abstract:  Mobile phones are attractive to use in studies of humans. Phones potentially allow researchers to study objective facts (like health) behaviors (time use, travel behavior) or attitudes (asking questions in the moment) in a better, faster or cheaper way than using other research methods, such as surveys. Mobile phones can be used to ask survey questions, but also allow location-tracking, the study of human interactions, the use of pictures, and various other sensors. Many current examples of app studies are being conducted in small samples or among volunteers however, and it often remains unclear whether the estimates from these studies are accurate and useful for policy-making. There are very few studies that try to generalize findings from their studies to the general population, let alone produce official statistics. Willingness, nonresponse, and the processing of data are potentially big hurdles in running smartphone-app studies at a large scale.

In this presentation I will report on a series of studies that are being carried out among the general population in the Netherlands to test whether smartphone apps can be used to replace surveys in producing official statistics on travel behavior, time use and household spending. I will present a framework that can be used to determine for which particular topics and populations it is worthwhile to consider using smartphones-studies. The framework is then used to assess in more detail two specific studies carried out in 2018 and 2019: 1. a smartphone-app to study travel behavior and 2. a browser-based study to understand more about people’s home environment. For both applications, we will discuss errors in non-participation, the quality of the data, data processing, and finally compare the quality of the estimates that are produced through sensor-based smartphone studies to the quality of estimates conducted with surveys alone. Are the smartphone-based studies really better than surveys?