What’s the noise? Understanding the Commerce Department’s directive on noise infusion

The U.S. Department of Commerce’s June 2026 directive restricting the use of noise infusion in federal statistical products has sparked heated debate across the statistical, research, and policymaking communities.

美国商务部 2026 年 6 月的指令限制在联邦统计产品中使用噪声注入，引发了统计、研究和政策制定界的激烈争论。

The policy represents one of the most consequential shifts in federal disclosure-avoidance practices in recent years, raising important questions about how government agencies will balance privacy protection, data accuracy, and public access to information.

该政策代表了近年来联邦避免披露做法中最重要的转变之一，引发了有关政府机构如何平衡隐私保护、数据准确性和公众获取信息的重要问题。

Why did the Commerce Department make this change now?

商务部为何现在做出这一改变？

Does restricting the use of noise infusion represent a political choice?

限制噪音输入的使用是否代表一种政治选择？

The directive states clearly that the goal is to provide the public with “accurate and objective information,” but the scientific community has generally endorsed noise infusion to provide detailed information across geographic areas and population subgroups without introducing disclosure risk.

该指令明确指出，目标是向公众提供“准确和客观的信息”，但科学界普遍支持噪声注入，以提供跨地理区域和人口亚组的详细信息，而不会引入披露风险。

Understanding the likely impact of the directive requires a deep dive into the tradeoffs associated with the various approaches to disclosure avoidance.

要了解该指令可能产生的影响，需要深入研究与各种避免披露方法相关的权衡。

Noise infusion refers to a class of disclosure-avoidance techniques that introduce small, carefully calibrated modifications into published statistics to protect the confidentiality of individuals and businesses while preserving the overall analytical value of the data.

噪声注入是指一类避免披露的技术，该技术对已发布的统计数据进行小的、经过仔细校准的修改，以保护个人和企业的机密性，同时保留数据的整体分析价值。

These methods, including differential privacy, have become increasingly important as advances in computing power, artificial intelligence, and data integration have made it easier to reconstruct confidential information from publicly available datasets.

随着计算能力、人工智能和数据集成的进步使得从公开数据集中重建机密信息变得更加容易，包括差异隐私在内的这些方法变得越来越重要。

Under the new directive, federal statistical agencies, including the U.S. Census Bureau and the Bureau of Economic Analysis, are directed to rely primarily on more traditional disclosure-limitation methods such as aggregation, suppression, rounding, and other forms of data coarsening.

根据新指令，包括美国人口普查局和经济分析局在内的联邦统计机构被要求主要依赖更传统的披露限制方法，例如汇总、抑制、舍入和其他形式的数据粗化。

Supporters view these approaches as more transparent and easier to understand.

支持者认为这些方法更加透明且更容易理解。

Critics, however, argue that the policy removes a key tool that agencies use to simultaneously protect confidentiality and maintain detailed, useful public statistics, and could even eliminate some public data releases .

然而，批评者认为，该政策取消了各机构用来同时保护机密和维护详细、有用的公共统计数据的关键工具，甚至可能取消一些公共数据的发布。

The significance of the commerce directive extends well beyond a technical debate about statistical methodology.

商业指令的重要性远远超出了有关统计方法的技术辩论。

At its core, the policy affects how federal statistical agencies will protect confidential information while continuing to publish the detailed economic and demographic data relied upon by policymakers, businesses, researchers, and the public.

该政策的核心是影响联邦统计机构如何保护机密信息，同时继续发布政策制定者、企业、研究人员和公众所依赖的详细经济和人口数据。

Many experts argue that the choice is not between noise-infused data and perfectly accurate data.

许多专家认为，选择不是在充满噪声的数据和完全准确的数据之间进行的。

Rather, they contend that eliminating noise infusion may force agencies to rely more heavily on data suppression, aggregation, and reduced geographic or demographic detail in order to meet statutory confidentiality requirements.

相反，他们认为，消除噪音输入可能会迫使各机构更加依赖数据抑制、聚合以及减少地理或人口统计细节，以满足法定的保密要求。

In that scenario, public data products could become less useful, even if they contain fewer noise-based adjustments.

在这种情况下，公共数据产品的用处可能会降低，即使它们包含的基于噪声的调整较少。

This concern has been echoed by researchers , economists , and privacy experts who argue that modern disclosure risks require modern privacy protections.

这种担忧得到了研究人员、经济学家和隐私专家的呼应，他们认为现代的信息披露风险需要现代的隐私保护。

As data-linkage capabilities continue to expand, traditional disclosure-avoidance methods may be insufficient to protect confidentiality while maintaining the level of detail users have come to expect from federal statistics.

随着数据链接能力的不断扩展，传统的避免披露方法可能不足以保护机密性，同时保持用户期望从联邦统计数据中获得的详细程度。

The importance of the directive was underscored by its prominence at the 2026 Joint Statistical Meetings (JSM), the largest gathering of statisticians in North America.

该指令的重要性在 2026 年联合统计会议 (JSM)（北美最大的统计学家聚会）上得到了突出体现。

An invited session, Updates on Disclosure Avoidance at the U.S. Census Bureau , brought together senior leaders from the Census Bureau to discuss the rationale behind the policy and its implications for major statistical programs.

一场名为“美国人口普查局避免披露的最新动态”的特邀会议汇聚了人口普查局的高级领导人，讨论该政策背后的基本原理及其对主要统计项目的影响。

The session’s extensive audience engagement reflected widespread concern about how agencies will balance privacy protection, data quality, and data accessibility moving forward.

会议的广泛观众参与反映了人们对各机构未来如何平衡隐私保护、数据质量和数据可访问性的广泛关注。

Several important clarifications also emerged during these discussions, and have been clarified by the Census Bureau in more detail .

在这些讨论中还出现了一些重要的澄清，并已由人口普查局进行了更详细的澄清。

The directive primarily affects official federal statistical releases.

该指令主要影响官方的联邦统计数据发布。

Research products distributed under appropriate disclosure protections, including certain Federal Statistical Research Data Center (FSRDC) projects and synthetic data products (e.g., the SIPP (Survey of Income and Program Participation) Synthetic Beta and validation process), remain exempt from the policy.

在适当的披露保护下分发的研究产品，包括某些联邦统计研究数据中心 (FSRDC) 项目和合成数据产品（例如 SIPP（收入和计划参与调查）合成 Beta 和验证过程），仍然不受该政策的约束。

Concerns about the directive extend beyond the statistical community.

对该指令的担忧超出了统计界的范围。

Analysts have warned that restricting modern disclosure-avoidance methods could reduce the availability of detailed data used for economic analysis, workforce research, tax policy evaluation, and program assessment.

分析师警告说，限制现代信息披露方法可能会减少用于经济分析、劳动力研究、税收政策评估和项目评估的详细数据的可用性。

Researchers argue that if agencies lose access to noise-based protections, they may be forced to publish less granular information, limiting insights into local communities, labor markets, and demographic trends.

研究人员认为，如果各机构无法获得基于噪音的保护，他们可能会被迫发布不太精细的信息，从而限制对当地社区、劳动力市场和人口趋势的了解。

Economic policy experts have similarly cautioned that reduced access to detailed public data could affect evidence-based decision-making by governments, businesses, and researchers.

经济政策专家同样警告说，获取详细公共数据的机会减少可能会影响政府、企业和研究人员基于证据的决策。

Rather than improving data quality, critics argue that the directive could ultimately result in less informative public statistics.

批评者认为，该指令最终可能会导致公共统计数据信息量减少，而不是提高数据质量。

The debate therefore centers not only on privacy methodology but also on the broader question of who gets access to information and how much detail can realistically be provided while protecting confidentiality.

因此，争论不仅集中在隐私方法上，还集中在更广泛的问题上，即谁可以访问信息以及在保护机密性的同时可以实际提供多少细节。

The policy has also attracted congressional attention.

该政策也引起了国会的关注。

Senator Elizabeth Warren , ranking member of the Senate Banking, Housing, and Urban Affairs Committee, questioned both the rationale for the directive and the process by which it was adopted.

参议院银行、住房和城市事务委员会资深成员、参议员伊丽莎白·沃伦对该指令的基本原理及其通过过程提出了质疑。

In correspondence with federal agencies, the senator raised concerns about whether the policy was supported by sufficient technical evidence, whether experts were adequately consulted, and whether limiting disclosure-avoidance options could reduce the usefulness of key economic and demographic statistics.

在与联邦机构的通信中，这位参议员对政策是否得到足够的技术证据支持、是否充分咨询专家以及限制披露避免选项是否会降低关键经济和人口统计数据的有用性表示担忧。

Their involvement elevates the issue beyond a methodological dispute within the statistical community and highlights its implications for government transparency, resource allocation, economic measurement, and public accountability.

他们的参与使这个问题超越了统计界内部的方法论争议，并突显了其对政府透明度、资源分配、经济衡量和公共问责制的影响。

The letter also reflects a broader concern that significant changes to federal statistical practices should be transparent, evidence-based, and informed by technical expertise.

这封信还反映了一种更广泛的担忧，即联邦统计实践的重大变化应该是透明的、基于证据的，并由技术专业知识提供信息。

The Commerce Department’s directive restricting noise infusion represents a significant shift in federal statistical policy.

商务部限制噪音输入的指令代表了联邦统计政策的重大转变。

While the policy is focused on disclosure-avoidance methods, its implications reach much further, affecting how federal agencies manage the longstanding tension between confidentiality, accuracy, and data utility.

虽然该政策的重点是避免披露的方法，但其影响更深远，影响联邦机构如何处理保密性、准确性和数据实用性之间长期存在的紧张关系。

The central concern raised by researchers, statisticians, and policymakers is not simply whether noise infusion should be used, but whether eliminating it will leave agencies with fewer options to protect privacy without reducing the detail and usefulness of public data.

研究人员、统计学家和政策制定者提出的主要担忧不仅仅是是否应该使用噪声注入，而是消除噪声注入是否会导致各机构在不降低公共数据的细节和有用性的情况下保护隐私的选择更少。

As agencies implement the directive and prepare for future statistical programs, including the 2030 Census, the outcome of this debate will help shape the future of federal statistics, privacy protection, and evidence-based policymaking in the United States.

随着各机构执行该指令并为未来的统计计划（包括 2030 年人口普查）做准备，这场辩论的结果将有助于塑造美国联邦统计、隐私保护和循证决策的未来。

The Census Bureau, in their recent blog , stated that they will “…regularly engage with stakeholders to shape our future products and policies to ensure our products meet your needs.” Hence, if you have concerns, follow updates listed here , send comments to elected officials, and watch for Federal Register notices.

人口普查局在最近的博客中表示，他们将“……定期与利益相关者合作，制定我们未来的产品和政策，以确保我们的产品满足您的需求。”因此，如果您有疑虑，请关注此处列出的更新，向民选官员发送评论，并留意《联邦公报》通知。