What can we ask respondents in terms of data donation?

We assess one possible implementation of energy data donation, using consumer‐grade smart‐meter dongles combined with respondent diaries. Participants installed a HomeWizard P1 dongle and kept an eight‐day diary on occupancy and appliance use. We processed 15‐minute exported meter readings (electricity import/export, gas; per‐phase power where available), derived quarter‐hourly consumption, and linked diaries for interpretation. Installation was largely straightforward; respondents valued real‐time feedback, but diaries were burdensome. Quarter‐hourly data reveal daily patterns and, together with diaries, allow recognition of some high‐load appliances. Disaggregation is limited when activities overlap, and hybrid heating requires contextual cues. In households with solar panels, combining net import with power readings helps to reveal masked background loads. Energy data donation via consumer‐grade dongles is feasible and acceptable but, at 15‐minute resolution, insufficient for automated device‐level statistics. We recommend (i) a privacy‐by‐design approach controlled by the National Statistical Institute, (ii) a rotating panel with onboarding and a short stabilisation period, and (iii) higher‐frequency sensing and human‐in‐the‐loop validation to reduce burden and enable scalable production statistics.

Bakker, J., Kompier, M.E., & Toepoel, V. (2026). What can we ask respondents in terms of data donation? The case of energy usage. Discussion paper, Statistics Netherlands, The Hague/Heerlen.