Crowdsourcing brings distributed participation into research, public administration, archival work, and paid digital labor. The statistics below show how large these efforts can become, how participants contribute, and how crowdwork fits into working lives. Measurement periods and source contexts vary, so the figures should be read as distinct snapshots rather than a single global total.
Contents
- Federal crowdsourcing and citizen science
- How large citizen-science programs are
- Public participation in archives
- The time and tools of observation
- Crowdwork prevalence and platform use
- Hours, demand, and household income
- Who performs crowdwork and why
Federal crowdsourcing and citizen science
The federal crowdsourcing and citizen science report examined 86 projects across 14 federal agencies during fiscal years 2017 and 2018. It found that 71% of those projects had at least one non-federal partner, while 33% involved more than one agency. These figures describe collaboration within a defined federal project sample; they are not a count of all crowdsourcing activity in the United States. (CitizenScience.gov report to Congress 2019)
The partnership figures show two different forms of distributed work. Non-federal participation was present in most projects, whereas cross-agency participation occurred in about one-third. That distinction matters for interpreting crowdsourcing: a project can draw on public or private partners without being jointly managed by several federal agencies.
CitizenScience.gov says its toolkit was developed with support and collaboration from over 25 federal agencies. Its community of practice includes hundreds of citizen-science practitioners and coordinators across government. Those figures indicate an institutional network around citizen science, rather than the number of individual contributors or completed tasks. (CitizenScience.gov toolkit credits; CitizenScience.gov about page)
The federal catalog lists 504 projects. Its agency-specific entries include 35 NASA projects, 78 NSF projects, 174 National Park Service projects, and 44 USGS projects. These catalog counts are useful for comparing the listed portfolios, but they should not be added to the 86-project report sample: the two measurements describe different collections and periods. (CitizenScience.gov catalog; CitizenScience.gov NASA catalog; CitizenScience.gov NSF catalog; CitizenScience.gov NPS catalog; CitizenScience.gov USGS catalog)
How large citizen-science programs are
NASA’s current citizen-science page says 42 NASA science projects are open to everyone. NASA also says that volunteers and amateurs have helped make thousands of scientific discoveries through its projects, and that more than 650 NASA citizen scientists have co-authored publications. The 42-project count and the publication figure describe NASA’s current public-facing program information, while “thousands” is an intentionally non-exact description of discoveries. (NASA Citizen Science page)
Zooniverse provides another scale marker. CitizenScience.gov says the platform has engaged 1.5 million registered users and supported over 70 online citizen-science projects since 2007. Registered users are not necessarily equivalent to active users in every project, and the project count is cumulative over the stated period. (CitizenScience.gov Zooniverse catalog entry)
The measurements can be compared compactly, but they represent different units:
| Program or catalog measure | Reported figure | What it measures |
|---|---|---|
| Federal catalog | 504 projects | Listed federal citizen-science projects |
| NASA catalog entry | 35 projects | Projects in the catalog entry |
| NSF catalog entry | 78 projects | Projects in the catalog entry |
| NPS catalog entry | 174 projects | Projects in the catalog entry |
| USGS catalog entry | 44 projects | Projects in the catalog entry |
| NASA current page | 42 projects | NASA science projects open to everyone |
| Zooniverse | 1.5 million users | Registered users |
| Zooniverse since 2007 | Over 70 projects | Supported online projects |
The table places project inventories beside a user count to clarify the scale difference. A project count describes opportunities or organizational units; a registered-user count describes accounts associated with a platform. Neither is a direct measure of completed contributions.
Public participation in archives
The National Archives illustrates how crowdsourcing can help process a very large documentary record. The Citizen Archivist case study says the National Archives has more than 12 billion pages of paper records alone. It also says the National Archives Catalog contains more than 2.5 million pages of records. These are separate scale statements about the broader paper holdings and the cataloged records available through the catalog. (Citizen Archivist case study)
During one week, a public transcription challenge produced 2,500 transcribed pages, and the public added 10,000 tags to National Archives records. In the same case study, NARA said its records had been incorporated into more than 4,000 Wikipedia articles, which received more than 1 billion page views annually. The page-view figure describes those articles’ annual audience, not the number of people who contributed archival work. (Citizen Archivist case study)
Seven months after an expansion, the public had contributed more than 20,000 pages of transcriptions and added 40,000 tags to NARA records. The later snapshot is broader than the single-week challenge, so the two sets of figures should not be treated as one continuous production rate. The case study also describes Citizen Archivist contributions since 2012 as millions of tags, metadata items, transcriptions, video subtitles, and digital images. (Citizen Archivist case study)
These archival figures show several kinds of contribution: transcription converts visual records into searchable text, tags add descriptive metadata, and digital images or subtitles expand access in other formats. Because the measurements count different outputs, a page, a tag, an article, and a page view are not interchangeable units of productivity.
The time and tools of observation
Nature’s Notebook asks volunteers to spend about 10 minutes observing a few times a week. That schedule gives the program a low-duration participation model: contributors provide repeated observations in short sessions rather than necessarily completing a single large task. The figure is a participation expectation, not a measured average of every volunteer’s actual time. (Nature’s Notebook case study)
The mPING app shows why the channel used for contribution can matter. During its launch period, only 3% of volunteer observations came through a web interface. The remaining observations therefore came through other channels in that measurement period, but the supplied figure does not specify a detailed breakdown of those channels. (mPING case study)
Equipment costs can also shape who is able to contribute. The Air Sensor Toolbox case study says air-quality sensors priced as little as $100 were commercially available for citizen-science use. “As little as” identifies a lower-end price point, not the typical cost of a complete project or monitoring setup. (Air Sensor Toolbox case study)
The Bureau of Land Management’s crowdsourcing action plan covers 245 million acres managed by the agency. That geography gives a sense of the physical scale at which distributed observation or public input may be relevant. It does not state how many participants, observations, or projects covered those acres. (BLM Crowdsourcing and Citizen Science Action Plan)
Crowdwork prevalence and platform use
Crowdsourcing also includes paid tasks performed through digital platforms. The ILO survey found that 0.5% of the workforce performed platform work as their main job in one OECD-style survey. A separate survey found that crowdwork prevalence ranged from 9% of respondents in the Netherlands in 2016 to 44% in the Czech Republic in 2019. These estimates come from different surveys, countries, and measurement years, so the range is not a trend line. (OECD Measuring Platform-Mediated Workers; OECD Handbook on Measuring Digital Platform Employment and Work)
The ILO findings describe a workforce that often uses multiple platforms. Almost half of respondents had worked on more than one platform in the previous month, 21% had worked on three or more different platforms, and 51% had worked on only one platform. These categories should be interpreted according to the survey’s comparison group and reference month; “only one” in this context does not mean that a worker had never used another platform previously. (ILO Digital Labour Platforms and the Future of Work)
Experience was also established for many respondents. Fifty-six percent had been performing crowdwork for more than a year, while 29% had more than three years of crowdwork experience in 2017. The two figures use different duration thresholds, so they describe overlapping but not identical groups. (ILO Digital Labour Platforms and the Future of Work)
Hours, demand, and household income
Workers averaged 24.5 hours per week doing crowdwork. Of those hours, 18.6 were paid work and 6.2 were unpaid work. The reported components are survey averages and should not be used to infer an individual worker’s schedule. The ILO also reported that workers spent 20 minutes on unpaid activities for every hour of paid crowdwork, emphasizing that task time is not the only time associated with platform work. (ILO Digital Labour Platforms and the Future of Work)
Demand for more work was high in the same research. Eighty-eight percent of respondents wanted to do more crowdwork, and on average they wanted 11.6 more hours per week. Fifty-eight percent said the availability of tasks was insufficient; an additional 17% said they did not find enough well-paying tasks. These responses distinguish a shortage of available tasks from dissatisfaction with the pay quality of available tasks. (ILO Digital Labour Platforms and the Future of Work)
Crowdwork’s relationship to household income varied by workers’ position. For primary-income crowdworkers, 59% of total income came from crowdwork. For non-primary crowdworkers, crowdwork income and main-job income each averaged 36% of household income. The two measures use different income bases, so they should not be directly ranked as equivalent shares. (ILO Digital Labour Platforms and the Future of Work)
Who performs crowdwork and why
The ILO comparison group included varied household situations: 61% of surveyed crowdworkers were married or cohabiting, 13% lived alone, and 43% had children living in their households. These percentages describe separate characteristics and are not intended to form a single exhaustive partition. (ILO Digital Labour Platforms and the Future of Work)
Educational attainment was also mixed. Thirty-seven percent of crowdworkers had a Bachelor’s degree, and 20% had a postgraduate degree or higher. The figures identify substantial participation by people with higher education, but they do not describe the educational distribution of all platform workers or the wider workforce. (ILO Digital Labour Platforms and the Future of Work)
The reasons given for crowdwork point to both income and flexibility. Thirty-two percent said the main reason was to complement pay from other jobs, while 22% said they preferred to work from home. Among women, 15% cited only being able to work from home, compared with 5% of men. The gender comparison is a reported difference in stated reasons, not a measure of the total share of women and men who perform crowdwork. (ILO Digital Labour Platforms and the Future of Work)