The Evolution and Interpretation of "Statistical Purposes"

Danielle C. Neiman; John L. Eltinge; Michael B. Hawes; Paul S. Marck; Sallie Ann Keller

REVIEW 2 major objections 5 minor 57 references

The Evolution and Interpretation of "Statistical Purposes"

T0 review · 2 major / 5 minor · reviewed 2026-07-14 · grok-4.5

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read Laws already define 'statistical purposes' by two criteria—aggregate public-benefit statistics and a hard ban on using data to harm respondents—and a broader operational definition is needed.

desk verdict Solid historical-legal synthesis of two statutory criteria for “statistical purposes,” plus a candid proposal for a broader NSO definition; modest novelty, clean framing, worth a referee. read the letter →

arxiv 2607.11778 v1 pith:ID2NSDCJ submitted 2026-07-13 stat.OT stat.AP

Michael B. Hawes , John L. Eltinge , Paul S. Marck , Danielle C. Neiman , Sallie Ann Keller This is my paper

classification stat.OTstat.AP

keywords statisticalpurposesconfidentialitypublicbenefitdisclosureavoidanceprivacyprotectiontrustdataqualityofficialstatistics

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

National statistical offices routinely tell people that data will be used only for 'statistical purposes,' yet the phrase is rarely defined clearly for the public and is poorly understood by respondents. A review of U.S. statutes, regulations, and agency practice shows that the term has long rested on two concrete criteria: statistics must describe relatively large population groups and serve a general public benefit, and identifiable information must never be used for legal, regulatory, or other actions that harm the people or organizations who supplied it. The same phrase also sits inside a larger web of scientific-integrity rules, professional ethics codes, and the four statutory responsibilities of federal statistical agencies. The authors therefore offer a broader working definition that makes those implicit obligations explicit—accuracy, relevance, objectivity, transparency about limitations, stewardship of the public’s trust, and avoidance of group or purely partisan harm—so that the phrase can serve as a more reliable guide for both public messaging and day-to-day decisions.

What carries the argument

The two-criterion statutory core (aggregate public-benefit statistics plus functional separation that bars use of identifiable data for administrative, regulatory, or enforcement action against the data subject), expanded into an explicit multi-part operational definition that also requires scientific integrity, transparency, and avoidance of group or purely private/partisan harm.

What would settle it

Empirical cognitive testing or legal review showing that either (a) the two-criteria reading is not in fact the predominant interpretation across current U.S. statutes and agency practice, or (b) adding the broader integrity and group-harm language produces no measurable change in respondent understanding, trust, or agency decision quality.

Watch

Extended reading notes

Core claim

Reviews of underlying laws and policies identify two predominant criteria for 'statistical purposes': (1) production of statistical information about relatively large population aggregates, with the intention of creating a general public benefit; and (2) protection of the confidentiality of data collected about data subjects and a related prohibition against the use of that information for legal or regulatory action against those individuals or organizations. A broader applied definition that also incorporates scientific integrity, transparency about limitations, and avoidance of group or partisan harm is therefore warranted for NSO decision-making.

Load-bearing premise

The claim assumes that ethical frameworks and the four fundamental responsibilities of statistical agencies can and should be treated as load-bearing parts of the operational meaning of 'statistical purposes,' even though those frameworks are not themselves the statutory definition of the term.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

The paper’s real contribution is a clean extraction of two criteria that actually sit in the statutes (CIPSEA 502(5)/(7), Title 13 §9, Evidence Act 3563) and a multi-sentence broader definition offered explicitly “for consideration.” That definition is the only new piece; everything else is careful restatement of Taft 1910, the 1929 Act, wartime breaches, Privacy Act/FIPPs, Belmont/Menlo, ASA ethics, and parallel international language.

What it does well: the historical arc is accurate and readable, the two-criteria claim is directly supported by the quoted text, and the authors never pretend the ethical frameworks are themselves statutory definitions of the term. They flag the need for empirical work on public understanding and keep the proposal provisional. Citations are on-point; self-cites to Eltinge are peripheral. No circularity, no invented entities beyond the proposed definition itself.

Soft spots are real but proportionate. The leap from Belmont/Menlo/FIPPs and the four Evidence Act responsibilities into an operational NSO definition is normative, not deduced; the paper itself labels it a proposal, so the leap is not load-bearing for the descriptive claim. The international section is thin (one-sentence glosses). No data, no math, no formal verification—none claimed. Impact stays inside official-statistics governance and respondent messaging.

This is for people who write NSO policy, disclosure standards, or respondent materials, and for anyone teaching the legal-ethical boundary of official statistics. It is not a methods paper. A serious editor should send it to peer review; the synthesis is careful enough and the definition is concrete enough to deserve referee time, even if the final version stays modest. I would cite the two-criteria extraction and the proposed definition if I were writing on statistical confidentiality or agency trust. Bring it to reading group if the group cares about data-governance language; otherwise skip.

Referee Report

2 major / 5 minor

Summary. The paper traces the historical and legal evolution of the phrase “for statistical purposes only” in U.S. federal statistics, from Taft’s 1910 census proclamation through the 1929 Census Act, wartime breaches, Title 13, the Privacy Act, CIPSEA (2002), and the Evidence Act (2018). It extracts two predominant statutory criteria: (1) production of aggregate statistical information intended for general public benefit, and (2) confidentiality protections that prohibit use of identifiable data for legal or regulatory action against data subjects. It then situates the term within scientific-integrity principles, ASA/ISI ethical guidelines, and the Belmont/Menlo/FIPPs frameworks, and proposes a broader applied definition (Section 6) that adds transparency, scientific integrity, and avoidance of group or partisan harm. The paper closes with calls for empirical research on public interpretation of the phrase.

Significance. If the two-criteria extraction and the proposed broader definition are adopted as working guidance, NSOs would gain a clearer operational and ethical anchor for communications with respondents, internal data-use decisions, and stewardship of administrative data. The historical synthesis is carefully documented with primary statutory citations and is of direct practical value to statistical agencies facing declining trust and response rates. The explicit framing of the broader definition as “for consideration,” together with the forward-looking research agenda in Section 7, makes the contribution usable without overclaiming normative necessity.

major comments (2)

Section 6 (and the abstract’s criterion (1)): the claim that “intention of creating a general public benefit” is one of the two predominant criteria extracted from underlying laws is only weakly supported by the statutory language actually quoted. CIPSEA §§502(5),(7) and 13 U.S.C. §9 emphasize aggregate description without identification and prohibition of non-statistical uses that affect rights or benefits; they do not themselves codify “public benefit.” The public-benefit language appears mainly in Taft’s 1910 proclamation and in selected international examples (UK ONS, Statistics NZ). The manuscript should either (a) qualify the first criterion as an implicit or historical rather than statutory element, or (b) supply additional U.S. statutory or OMB policy language that explicitly embeds public benefit inside the definition of statistical purpose.
Sections 4 and 6: the leap from the Evidence Act’s four fundamental responsibilities (44 U.S.C. §3563) and the Belmont/Menlo/FIPPs frameworks into the multi-sentence “broader applied definition” is presented as a synthesis “for consideration,” yet the abstract and introduction still frame the paper as showing that reviews of laws and policies “identify” the two criteria and then “provide a broader definition.” To keep the descriptive claim load-bearing and the normative proposal clearly optional, the authors should add an explicit demarcation sentence at the start of Section 6 stating that the expanded definition is a policy recommendation, not a restatement of existing statutory text.

minor comments (5)

Section 2.1: the discussion of wartime uses of census data (Japanese-American internment, Second War Powers Act) is historically important but could be tightened; a short table or timeline of key statutory changes would help readers track the evolution of the confidentiality guarantee.
Section 5: the international comparison is limited to brief definitional quotations. A single comparative table (EUROSTAT, Statistics Canada, UNSD, UK ONS, Statistics NZ) would make the claimed “broad similarity” more transparent.
References: Eltinge (2025, 2026) are cited for related discussion; if these are forthcoming or under review, a note on availability would assist readers.
Typographical: “RELEV ANT” in the Evidence Act quotation (Section 4) contains a stray space; “oﬀicial” appears with a ligature that may render inconsistently.
Section 7: the proposed empirical research agenda is valuable; specifying one or two concrete cognitive-interview or survey-experiment designs would strengthen the call for future work.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: historical-legal synthesis and an explicitly labeled proposal, not a derivation that reduces to its inputs.

full rationale

This paper is a historical, legal, and ethical review of the phrase “statistical purposes,” not a quantitative derivation. Its central descriptive claim—that reviews of U.S. law and policy identify two predominant criteria (aggregate public-benefit statistics and confidentiality / non-enforcement use)—is grounded in external primary sources (Taft 1910 proclamation, 1929 Census Act, Title 13, Privacy Act, CIPSEA §502, Evidence Act 44 U.S.C. §3563, etc.). The broader applied definition in Section 6 is explicitly framed as a synthesis “for consideration,” not as a theorem deduced from itself. Self-citations (Eltinge 2025, 2026) appear only as peripheral pointers to related discussion and are not load-bearing premises for either the two-criteria extraction or the proposed definition. There are no fitted parameters, no uniqueness theorems imported from the authors, no ansatz smuggled via self-citation, and no renaming of a known result presented as a new derivation. The paper is self-contained against its external statutory and ethical benchmarks; circularity score is therefore 0.

Assumptions & free parameters 0 free parameters · 4 assumptions · 1 invented entities

The paper is a legal-historical and ethical synthesis; it rests on domain assumptions drawn from U.S. statutes, OMB directives, and classic research-ethics reports rather than free parameters or newly postulated physical entities. No numerical fitting occurs.

assumptions (4)

domain assumption CIPSEA (as amended by the Evidence Act) defines statistical purpose as description/estimation/analysis of group characteristics without identifying individuals and excludes administrative, regulatory, or law-enforcement uses that affect rights or benefits of identifiable respondents.
Invoked throughout Sections 2.3 and 4 as the current statutory baseline.
domain assumption The four fundamental responsibilities of a federal statistical agency (relevance, credibility/accuracy, objectivity, confidentiality/trust) codified at 44 U.S.C. §3563(a) correctly capture the operational duties of NSOs.
Used in Section 4 to expand the meaning of statistical purpose beyond the 2002 CIPSEA text.
domain assumption The Belmont Report principles (respect for persons, beneficence, justice) and the Menlo Report extensions (stakeholder perspectives, public interest, transparency, accountability) supply ethically binding guidance for the stewardship of data-subject information by statistical agencies.
Section 6 treats these frameworks as normative inputs to the broader definition.
domain assumption Declining public trust and survey response rates create a practical need for clearer purpose-specification language.
Section 3 cites Pew and response-rate series as background motivation; the causal link to the definitional proposal is assumed rather than demonstrated.

invented entities (1)

Broader applied definition of “statistical purposes” (Section 6 multi-sentence formulation)
purpose: To make explicit the implicit public-benefit, integrity, transparency, and harm-avoidance elements that the authors argue should guide NSO decisions.
The definition is newly composed by the authors; it is offered for consideration rather than derived from a uniqueness theorem or empirical fit.

how reviews work

0 comments

Cite this review

Pith. "Pith review of The Evolution and Interpretation of "Statistical Purposes"." pith.science (2026). https://pith.science/paper/ID2NSDCJ

@misc{pith2026260711778,
  author       = {Pith},
  title        = {Pith review of: The Evolution and Interpretation of "Statistical Purposes"},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ID2NSDCJ}},
  note         = {Machine review of arXiv:2607.11778}
}

read the original abstract

National Statistical Organizations (NSOs) and other groups often use the term "for statistical purposes only" in communication with prospective respondents, data users and other stakeholders. This term also provides an important anchor for many NSO decisions on operations and ethics. Although public communication often omits a clear operational definition of this term, this paper will show that reviews of underlying laws and policies identify two predominant criteria: (1) production of statistical information about relatively large population aggregates, with the intention of creating a general public benefit; and (2) protection of the confidentiality of data collected about data subjects and a related prohibition against the use of information provided by or about data subjects for legal or regulatory action against those individuals or organizations. We then explore how this term exists, and is often interpreted, within a much broader landscape of legal requirements and of scientific and professional codes of practice, and provide a broader definition for consideration based on these and other ethical frameworks. This paper closes by highlighting several areas that warrant further discussion to better position NSOs to navigate these challenges in the future.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

57 extracted references · 2 canonical work pages

[1]

arXiv preprint arXiv:2502.07105 , year=

Towards a Principled Framework for Disclosure Avoidance , author=. arXiv preprint arXiv:2502.07105 , year=

arXiv
[2]

Our Privacy Principles , year = 2006, url =

2006
[3]

OImport/Export Price Indexes (MXP) Survey Respondents: Confidentiality of Data Collected by BLS for Statistical Purposes , year = 2019, url =

2019
[4]

National Electronic Health Records Survey: NEHRS Participants , year = 2024, url =

2024
[5]

Statistical Research Division Report Series (Survey Methodology\# 2008-4)

Report of Cognitive Testing of Privacy and Confidentiality-Related Statements in Respondent Materials for the 2010 Decennial: Results from Cognitive Interview Pretesting with Volunteer Respondents.” US Census Bureau , author=. Statistical Research Division Report Series (Survey Methodology\# 2008-4). On-line, available: http://www. census. gov/srd/papers/...

2010
[6]

Statistical Standard 4-2: Maintaining Confidentiality , year = 2002, url =

2002
[7]

The Federal Statistical System , year = 2025, url =

2025
[8]

Stanford Encyclopedia of Philosophy , year=

Public goods , author=. Stanford Encyclopedia of Philosophy , year=

Show all 57 references

[9]

Statistics , journal =
[10]

, title =

Taft, William H. , title =
[11]

US Bureau of the Census, Washington , year=

A monograph on confidentiality and privacy in the US Census , author=. US Bureau of the Census, Washington , year=
[12]

Presentation at the Population Association of America Annual Meeting , year=

Census confidentiality under the second war powers act (1942-1947) , author=. Presentation at the Population Association of America Annual Meeting , year=

1942
[13]

Census Bureau , year = 2002, url =

Events in the Chronological Development of Privacy and Confidentiality at the U.S. Census Bureau , year = 2002, url =

2002
[14]

1977 , publisher=

Personal Privacy in an Information Society: The Report of the Privacy Protection Study Commission , author=. 1977 , publisher=

1977
[15]

Public Trust in Government: 1958-2024 , year = 2025, url =

1958
[16]

Survey Methods: Insights from the Field , pages=

Trust and credibility in the US Federal Statistical System , author=. Survey Methods: Insights from the Field , pages=. 2019 , publisher=

2019
[17]

Mathematica policy research , volume=

Background paper declining response rates in federal surveys: Trends and implications , author=. Mathematica policy research , volume=
[18]

American Community Survey Response Rates , year = 2024, url =

2024
[19]

2015 , publisher=

Americans’ attitudes about privacy, security and surveillance , author=. 2015 , publisher=

2015
[20]

International journal of Public Opinion research , volume=

Confidentiality assurances in surveys: Reassurance or threat? , author=. International journal of Public Opinion research , volume=. 1992 , publisher=

1992
[21]

2001 , publisher=

SIPP Advance Letter Research: Cognitive Interv [i] ew Results, Implications, & Letter Recommendations , author=. 2001 , publisher=

2001
[22]

Final Report, Census , year=

Privacy Schemas and Data Collection: An Ethnographic Account , author=. Final Report, Census , year=
[23]

The Promise of Evidence-based Policymaking: Report of the Commission on Evidence-based Policymaking , year = 2017, url =

2017
[24]

Statement of Commitment to Scientific Integrity by Principle Statistical Agencies , year = 2012, url =

2012
[25]

Census Bureau Scientific Integrity Policy , year = 2024, url =

2024
[26]

Ethical Guidelines for Statistical Practice , year = 2022, url =

2022
[27]

Declaration on Professional Ethics , year = 2010, url =

2010
[28]

Statistical Quality Standards , year = 2023, url =

2023
[29]

Statistical requirements compendium , year = 2023, url =

2023
[30]

Statistics Canada Quality Assurance Framework , year = 2017, url =

2017
[31]

United Nations National Quality Assurance Frameworks Manual for Official Statistics , year = 2019, url =

2019
[32]

Data Protection: Privacy Information , year = 2023, url =

2023
[33]

Wisconsin Public Radio , year =

Hansi Lo Wang , title =. Wisconsin Public Radio , year =
[34]

2024 , month = feb, day =

Hansi Lo Wang , title =. 2024 , month = feb, day =

2024
[35]

2023 , institution=

Do No Harm Guide: Applying Equity Awareness in Data Privacy Methods , author=. 2023 , institution=

2023
[36]

2019 , month = nov, day =

Andrew Reamer , title =. 2019 , month = nov, day =

2019
[37]

Statistical Journal of the IAOS , volume=

Statistics for the public good: What it means and why it matters , author=. Statistical Journal of the IAOS , volume=. 2024 , publisher=

2024
[38]

doi:10.17226/27934 , url =

Principles and Practices for a Federal Statistical Agency: Eighth Edition , isbn =. doi:10.17226/27934 , url =

doi
[39]

2020 , month = sep, url =

2020
[40]

Survey methodology , volume=

Managing data quality in a statistical agency , author=. Survey methodology , volume=. 1999 , publisher=

1999
[41]

Journal of Official Statistics , volume=

Some Dimensions of Statistical Ethics and Scientific Integrity That Warrant Exploration Through Empirical Studies of Stakeholder Information Needs , author=. Journal of Official Statistics , volume=. 2025 , publisher=

2025
[42]

Data Science in Science , volume=

Ethical Issues in the Dynamic Environment for Public-Stewardship Statistical Information , author=. Data Science in Science , volume=. 2026 , publisher=

2026
[43]

2014 , publisher=

Hard-to-survey populations , author=. 2014 , publisher=

2014
[44]

2015 , publisher=

Small area estimation , author=. 2015 , publisher=

2015
[45]

Clark, Cynthia Z. F. , title =. 2020 , month = mar, note =

2020
[46]

2022 , howpublished =

2022
[47]

2025 , month = may, howpublished =

2025
[48]

, title =

Villa Ross, Ceci A. , title =. 2023 , month = jun, type =

2023
[49]

1979 , month = apr, day =

1979
[50]

1973 , month = jul, url =

1973
[51]

Organizing to Count: Change in the Federal Statistical System , author=

Norwood. Organizing to Count: Change in the Federal Statistical System , author=. 1995 , publisher=

1995
[52]

The American Statistician , volume=

Report of the Task Force on the Storage of and Access to Government Statistics , author=. The American Statistician , volume=. 1969 , publisher=

1969
[53]

1973 , publisher=

Records, Computers, and the Rights of Citizens: Report , author=. 1973 , publisher=

1973
[54]

2012 , institution=

The Menlo Report: Ethical principles guiding information and communication technology research , author=. 2012 , institution=

2012
[55]

, title =

Lindley, Dennis V. , title =. Journal of the Royal Statistical Society: Series D (The Statistician) , volume =. doi:https://doi.org/10.1111/1467-9884.00238 , url =. https://rss.onlinelibrary.wiley.com/doi/pdf/10.1111/1467-9884.00238 , abstract =

doi
[56]

2014 , month = feb, day =

2014
[57]

Scientific data , volume=

The FAIR Guiding Principles for scientific data management and stewardship , author=. Scientific data , volume=. 2016 , publisher=

2016

Pith tools

Reviewed July 14, 2026 · model on record in the stance chip above.

[1] [1]

arXiv preprint arXiv:2502.07105 , year=

Towards a Principled Framework for Disclosure Avoidance , author=. arXiv preprint arXiv:2502.07105 , year=

arXiv

[2] [2]

Our Privacy Principles , year = 2006, url =

2006

[3] [3]

OImport/Export Price Indexes (MXP) Survey Respondents: Confidentiality of Data Collected by BLS for Statistical Purposes , year = 2019, url =

2019

[4] [4]

National Electronic Health Records Survey: NEHRS Participants , year = 2024, url =

2024

[5] [5]

Statistical Research Division Report Series (Survey Methodology\# 2008-4)

Report of Cognitive Testing of Privacy and Confidentiality-Related Statements in Respondent Materials for the 2010 Decennial: Results from Cognitive Interview Pretesting with Volunteer Respondents.” US Census Bureau , author=. Statistical Research Division Report Series (Survey Methodology\# 2008-4). On-line, available: http://www. census. gov/srd/papers/...

2010

[6] [6]

Statistical Standard 4-2: Maintaining Confidentiality , year = 2002, url =

2002

[7] [7]

The Federal Statistical System , year = 2025, url =

2025

[8] [8]

Stanford Encyclopedia of Philosophy , year=

Public goods , author=. Stanford Encyclopedia of Philosophy , year=