Data collection is both a practical infrastructure and a source of public concern. Recent figures show how people respond to surveys, consent to tracking, share sensitive information and judge the organizations that collect or use it. They also show the operational scale of government surveys, fraud reporting, app oversight and commercial data brokerage.
Contents
- Public trust and privacy concerns
- How people respond to data collection
- Survey response rates and nonresponse
- Government and public-sector data use
- Fraud reports and the value of collected data
- Platform oversight and transparency
- Data brokers and the scale of aggregation
Public trust and privacy concerns
The Pew Research Center reported on October 18, 2023, that 71% of U.S. adults were very or somewhat concerned about how the government uses data collected about them. That was up from 64% in 2019. The same research found a wide gap between confidence in personal judgment and confidence in institutions: 78% said they trusted themselves to make the right decisions to protect personal information, while 71% were skeptical that anything they did would make much difference.
Among adults who had heard of artificial intelligence, 70% had little or no trust in companies to make responsible decisions about how they use AI in products. Concern about downstream use was even higher. In that group, 81% said information companies collect would be used in ways people are not comfortable with, and 80% said it would be used in ways not originally intended. At the same time, 62% said companies using AI to analyze personal details could make life easier.
These results describe a mixed attitude rather than simple rejection. People can see benefits in analysis while expecting weak control over how information is reused. Pew also found that 61% thought privacy policies were ineffective at explaining how companies use people’s data. That helps put consent behavior in context: 56% said they always, almost always or often click “agree” without reading privacy policies, 22% said they sometimes do so, and 18% said they rarely or never do so.
Trust also varies by institution and political affiliation. Pew reported that 77% had little or no trust in social media leaders to publicly admit mistakes and take responsibility for data misuse. Another 71% had little to no trust that technology leaders would be held accountable for missteps. On regulation, 78% of Democrats and 68% of Republicans said there should be more government regulation of what companies can do with customers’ personal information.
Children’s data produced particularly high concern. The Pew figures show that 89% were concerned about social media sites knowing personal information about children, 85% were concerned about advertisers using children’s online activity to target ads, and 84% were concerned about online games tracking children on the internet.
How people respond to data collection
Security burdens are part of the collection experience. Pew found that 71% of adults felt overwhelmed by the number of passwords they had to remember, while 45% felt anxious about whether their passwords were strong and secure. The reported consequences were substantial: 26% said someone had put fraudulent charges on their debit or credit card in the previous 12 months, 11% said someone had taken over their email or social media account without permission, and 7% said someone had attempted to open a line of credit or apply for a loan using their name.
Overall, 34% of Americans experienced at least one of those three identity or fraud issues during the previous year, according to the October 18, 2023, Pew Research Center figures. The measures are not a single count of unique incidents; they describe the share reporting at least one of the specified experiences.
Survey evidence from England shows that willingness to share information depends strongly on the expected benefit and the recipient. In the GOV.UK Participation Survey 2024/25, 46% agreed and 26% strongly agreed that they would share medical history or allergies with the NHS if it helped develop new medicines or treatments. For businesses pursuing the same purpose, 39% agreed and 17% strongly agreed.
Spending data showed a similar tradeoff. In the same survey, 40% would share spending-habits data if doing so provided access to cost-saving services, while 28% would not share it for that purpose. Cookie choices were more divided: 50% of adults in England accepted all cookies when a cookie pop-up appeared, 28% rejected all cookies, and 19% customized their choices. Only 33% agreed that cookie pop-ups help them control access to their personal data.
Survey response rates and nonresponse
Response rates show how much of a data-collection system successfully produces usable participation. The U.S. Census Bureau’s ACS housing-unit response rate table reports the following percentages:
| Year | Housing-unit response rate |
|---|---|
| 2020 | 71.2% |
| 2021 | 85.3% |
| 2022 | 84.4% |
| 2023 | 84.7% |
| 2024 | 82.9% |
The figures are measurement-year rates, not forecasts. The ACS housing-unit noninterview table for 2024 identifies several reasons a housing unit did not produce an interview: refusals were 9.5%, contact attempts halted were 3.6%, no one was home was 1.7%, other reasons were 1.1%, insufficient data was 0.8%, language problems were 0.3%, unable to locate was 0.1%, and temporarily absent was 0.1%. These categories describe the reported noninterview distribution and should not be read as a ranking of all possible barriers outside the table.
For group-quarters people, the U.S. Census Bureau reported response rates of 74.8% in 2021, 81.2% in 2022, 84.3% in 2023 and 80.6% in 2024. The 2024 group-quarters person noninterview table recorded 5.7% as GQ person refusal, 0.5% as unable to locate, 0.3% as resident temporarily absent, 0.7% as insufficient data, and 3.9% as GQ person other. Whole-group-quarters outcomes included 6.6% refusal and 1.8% other.
The U.S. Census Bureau’s Annual Survey of Public Pensions reported an overall response rate of 77.0% and a total quantity response rate of 99.1% for Total Cash and Investments. In the historical response table, unit response rate (URR) and total quantity response rate (TQRR) were 64% and 99% in 2017, 68% and 97% in 2018, 66% and 99% in 2019, 67% and 97% in 2020, 70% and 98% in 2021, 66% and 99% in 2022, and 82% and 99% in 2023. URR and TQRR measure different aspects of survey completeness, so they should not be treated as interchangeable.
Government and public-sector data use
The GOV.UK Participation Survey 2024/25 measured comfort with data being used to uncover patterns and trends. In England, 69% of adults were comfortable with the public sector doing so, compared with 64% for government and 45% for businesses. For using data to make better decisions and deliver services, comfort was 74% for the public sector, 72% for government and 70% for businesses.
The survey’s collection process also has measurable scale. A total of 34,378 respondents took part in 2024/25: 29,912 completed online and 4,466 completed on paper. The reported mode shares were 87% online and 13% paper. The compliance cost was 15,689 hours, 44 minutes and 24 seconds.
These results distinguish acceptance of a purpose from trust in a particular institution. Adults in England were more comfortable with data being used to improve decisions and services than with its use to uncover patterns and trends by businesses. That distinction matters for collaborative projects, where participants may support a public benefit but still want clear boundaries around reuse.
Fraud reports and the value of collected data
The Federal Trade Commission’s Consumer Sentinel Network Data Book 2024 reported 6.5 million consumer reports received in 2024. The FTC said Consumer Sentinel has collected tens of millions of consumer reports since 1997. These are reports received by the network, not a complete count of every fraud event.
The FTC also reported that consumers lost more than $12.5 billion to fraud in 2024, a 25% increase over the prior year. Among people who reported a fraud, 38% said they lost money in 2024, compared with 27% in 2023. Investment scams accounted for $5.7 billion in reported losses, up 24% from 2023, while imposter scams accounted for $2.95 billion.
The figures show why reporting systems matter to data collection. A report can document an individual experience, support trend analysis and expose high-loss categories, but reported totals remain different from the total amount of harm occurring in the population.
Platform oversight and transparency
Apple’s 2024 App Store Transparency Report said the App Store operated in 175 countries and regions and that every app was reviewed for compliance with App Review Guidelines. Apple reported preventing $2,023,389,488 in fraudulent transactions in 2024.
The platform’s weekly averages illustrate the volume of activity that oversight systems must process: 813,110,348 App Store visitors, 839,266,915 app downloads, 1,898,240,007 app redownloads, 66,022,387,774 automatic app updates, 638,006,162 manual app updates and 440,896,463 customer accounts searching the App Store. An average of 169,235 weekly apps appeared in the top 10 results of at least 1,000 searches, while 1,377,468 total apps appeared in those results.
Government requests and appeals add another layer. In 2024, Apple removed 1,307 apps from mainland China, 171 from Russia, 79 from South Korea, 55 from Ukraine and 50 from Jordan under government takedown demands. Apple received 26,224 appeals of app removals and restored 421 apps after appeals. It also received 17,715 account-preservation requests and preserved data for 43,991 accounts.
Apple’s transparency reporting distinguishes bounded estimates from exact totals. For 2025 H1, the U.S. national-security report listed 500–999 FISA non-content requests affecting 62,500–62,999 users or accounts, and 500–999 FISA content requests affecting 75,500–75,999 users or accounts. For 2024 H2, it listed 500–999 non-content requests affecting 55,000–55,499 users or accounts and 500–999 content requests affecting 76,000–76,499 users or accounts.
Microsoft’s Government Requests Report said it receives about 50,000 law-enforcement demands each year, affecting roughly 110,000 individual accounts. Less than 0.7% of its legal demands seek enterprise customer data. In the first half of 2024, Microsoft received 166 enterprise-related law-enforcement requests; 118 were rejected, withdrawn, found to have no data or redirected, while 48 compelled disclosure. Microsoft disclosed customer content in 27 of those requests and non-content data in 21, including content data in 16 cases to U.S. law enforcement within that set.
Data brokers and the scale of aggregation
The FTC’s Data Brokers: A Call for Transparency and Accountability reported that the nine data brokers studied obtained most of their data from other data brokers rather than directly from original sources. Seven of the nine provided data to each other. The study also documented the scale of aggregation: one broker held information on 1.4 billion consumer transactions and more than 700 billion aggregated data elements; another covered one trillion dollars in consumer transactions; and another added three billion new records each month.
The breadth of segmentation was also quantified. One studied broker had 3,000 data segments for nearly every U.S. consumer, while another obtained consumers’ contact information from 20 different sources. Consumer control was limited in the study: only one of the nine brokers allowed consumers to correct their information, and none allowed consumers to opt out of the use or sharing of personal information in risk-mitigation products.
Together, these data collection statistics point to a recurring governance challenge: collection can be extensive, useful and operationally sophisticated while still leaving people uncertain about purpose, accuracy, consent and accountability. Response rates, reporting volumes and transparency disclosures make the system measurable; trust and control determine whether people view participation as legitimate.