The dual nature of trust in participatory science: An investigation into data quality and household privacy preferences

There is a duality of trust in participatory science (citizen science) projects in which the data produced by volunteers must be trusted by the scientific community and participants must trust the scientists who lead projects. Facilitator organizations can diversify recruitment and broaden learning outcomes. We investigated the degree to which they can broker trust in participatory science projects. In Crowd the Tap, we recruited participants through partnerships with facilitators, including high schools, faith communities, universities, and a corporate volunteer program. We compared data quality (a proxy for scientists’ trust in the project) and participant privacy preferences (a proxy for participants’ trust in the project leaders) across the various facilitators as well as to those who came to the project independently (unfacilitated). In general, we found that data quality differed based on the project’s level of investment in the facilitation partner in terms of both time and money. We also found that demographic characteristics, rather than facilitation, was most important in predicting privacy preferences. Ultimately our results reveal several tradeoffs that project leaders and facilitators should weigh when deciding to work together.

Methods
The data was collected through an IRB approved survey in which Crowd the Tap participants submitted data on the types of pipes they had, the age of their home, water aesthetics, and demographic information. As part of this process, participants also indicated if they came to the project through a partner organization (what we call facilitator organizations). Using the data available to us, we determined how completely, accurately, and informatively (understandability) they participated in the project to assess data quality. We also asked if they had interest in being publically associated with the project or if they referred to remain private. We used this and the number of times they selected "Prefer not to say" as indicators of privacy. We compared data quality and privacy preferences to the facilitator organization through which they came to the project.