Research
Across the three lines of research the question is the same: which part of experimental control is lost at each step from the laboratory to the browser, the smartphone and real-time events, and how it can be measured, restored or taken into account in the analysis.
Ongoing work extends the question to data quality when online participants may be AI agents, and to the measurement properties of smartphone sensors.
In an online experiment the procedure runs in the participant’s own browser, without an experimenter present. The researcher cannot see whether instructions were read, whether a video used for a mood induction was watched, or on which device and screen the stimuli appeared.
Experience sampling asks people about their experience and behaviour several times a day, in the situations in which they occur. Which situations end up in the data depends on when the prompts arrive: on a fixed or random schedule, or when an event such as arriving at a place occurs.
When the study material itself is sent to the phone, the researcher also gives up control over how stimuli are presented and how responses are recorded. Screen size, input method and the moment of delivery differ between participants and over the course of a day.