Headline points · The spectrum of self-reporting is extremely diverse and impossible to govern through a one size fits all approach; projects can be scientific or not, can raise important privacy concerns, and can be open to organised manipulation and poor governance of risks and harms; best practices and guidelines must be adapted to the local setting and remain open for improvement · People take part in data collection for the most heterogeneous of purposes. Closely controlling hopes, aims and beliefs of participants is beyond the reach of any sensible self-reporting exercise; this heterogeneity introduces various biases in the data, and adds unpredictability to the data collection project · Governance of self-report data collection is a thorny issue; new data governance approaches have been put forth to deal with the heterogeneity of ends that big data collection projects now need to be able to support, that require some experimentation, but the best chance of long-term sustainable governance comes from collective governance frameworks · Greatest scientific potential of self-reporting is through data linkage with other forms of heterogeneous data, however these forms of data linkage are also the source of the highest risks for privacy and individual harms; flexible and sustained ethical oversight is key · Participation in self-reporting is not equally distributed across society; when there are benefits associated in participation, it can exacerbate existing inequalities