Citizen Science Projects: How the Public Contributes to Research
Citizen science is the practice of enlisting non-professional volunteers to collect, classify, or analyze data as part of legitimate scientific research. It spans disciplines from astronomy to epidemiology, and the scale of participation is genuinely striking — the SciStarter project database lists over 3,000 active citizen science projects available to the public. This page covers how these projects are structured, what kinds of tasks volunteers actually perform, and how scientists decide where amateur contribution is scientifically valid and where it isn't.
Definition and scope
Citizen science sits at the intersection of public engagement and research methodology. At its core, it is a model in which trained or lightly trained non-experts contribute labor — observational, computational, or analytical — to projects that professional researchers design and oversee.
The phrase gained institutional traction in the 1990s, attributed simultaneously to ornithologist Rick Bonney at the Cornell Lab of Ornithology and sociologist Alan Irwin. Cornell's work with bird population monitoring programs — including the Christmas Bird Count, which the Audubon Society has run continuously since 1900 — demonstrated that aggregated volunteer data could meet publication-quality standards when protocols were tight enough.
The scope today is broad. NASA, NOAA, the U.S. Geological Survey, and the National Institutes of Health all maintain or fund citizen science programs. The federal CitizenScience.gov portal, managed by the General Services Administration, catalogs over 500 federally affiliated projects as of the platform's published providers. Understanding how science operates at this scale — how hypotheses become protocols become datasets — is foundational context; the how-science-works-conceptual-overview page covers that broader framework.
How it works
Most citizen science projects follow a five-stage structure:
- Protocol design — Researchers define exactly what volunteers will observe or do, and under what conditions. This step determines data quality more than any other.
- Volunteer recruitment and training — Participants receive instructions, often through a digital platform, that standardize how observations are recorded.
- Data collection — Volunteers submit data via apps, web forms, or direct uploads. Projects like iNaturalist receive millions of observations annually.
- Validation and quality control — Submitted data is filtered for outliers, cross-checked against expert observations, or subjected to consensus algorithms (where multiple volunteers classify the same item independently).
- Analysis and publication — Researchers integrate the cleaned dataset into formal analysis, which may then be submitted to peer-reviewed journals.
The validation step is where citizen science earns or loses its scientific credibility. Galaxy Zoo, launched in 2007 through the Zooniverse platform, demonstrated that when 10 or more volunteers independently classify the same galaxy image, the consensus accuracy rate approaches that of professional astronomers. That project has contributed to over 50 peer-reviewed publications according to Zooniverse's published project records.
The Science Funding and Grants page details how projects like these secure institutional backing — because most citizen science programs require sustained infrastructure even when the labor is free.
Common scenarios
Citizen science clusters into three broad types by task structure:
Contributory projects give volunteers a defined task — count birds, photograph plants, record weather observations — with no input into study design. The majority of large-scale projects fall here. eBird, run by the Cornell Lab of Ornithology, has accumulated over 1 billion bird observation records from volunteers worldwide (Cornell Lab of Ornithology, eBird).
Collaborative projects involve volunteers in data interpretation or problem-solving. The SETI@home project, which ran from 1999 to 2020 through UC Berkeley's Space Sciences Laboratory, used volunteer computing power to analyze radio telescope data — a hybrid of contributory data processing and genuinely distributed scientific labor.
Co-created projects allow community members to participate in defining the research question itself. These are rarer and tend to emerge in environmental health contexts, where affected communities have both stake in the outcome and localized knowledge professionals lack.
Across all three types, the most productive domains share a common trait: the task can be broken into discrete, repeatable units that don't require deep disciplinary expertise to execute correctly.
Decision boundaries
Not every research question is suited to citizen science, and conflating enthusiasm with methodological fitness is one of the field's recurring problems. The science reference portal at /index covers the broader landscape of scientific domains — citizen science intersects with some of those fields cleanly and with others only awkwardly.
Citizen science works reliably when:
It performs poorly when:
The academic literature on this tension is substantial. A 2016 review published in Biological Conservation (Aceves-Bueno et al.) found that citizen science data quality was "comparable to professional data" in 79% of studies reviewed — but that the remaining 21% showed systematic biases traceable to volunteer recruitment patterns or protocol ambiguity, not volunteer incompetence. The distinction matters: the tool isn't broken, but it requires honest appraisal of where it applies.