ASA Ethical Guidelines for Statistical Practice: Guide

The ASA Ethical Guidelines for Statistical Practice are the American Statistical Association’s code for people who design, analyze, interpret and present data, or who develop and deploy models and algorithms. They were prepared by the ASA Committee on Professional Ethics and approved by the ASA Board of Directors on February 1, 2022. The guidelines are written for individuals, but the ASA states that the principles are also relevant to organizations that engage in statistical practice.

Within statistical research, the guidelines frame ethical practice as concrete responsibilities that run from designing data collection through to publication and model deployment.

What Are the ASA Ethical Guidelines for Statistical Practice?

The ASA says the guidelines are intended to help statistical practitioners make decisions ethically. They define “statistical practice” to include designing the collection of data, summarizing, processing, analyzing, interpreting or presenting data, and model or algorithm development and deployment. The term “statistical practitioner” covers everyone who engages in statistical practice, regardless of job title, profession, level or field of degree.

The guidelines are organized into eight principles, followed by an appendix on the responsibilities of organizations and institutions:

• Principle A: Professional Integrity and Accountability. Taking responsibility for one’s work, representing capabilities honestly, disclosing conflicts of interest and resisting pressure to selectively interpret data.
• Principle B: Integrity of Data and Methods. Understanding and mitigating limitations and biases in data and methods, communicating data sources, processing steps and assumptions, and being transparent about planned versus unplanned analyses.
• Principle C: Responsibilities to Stakeholders. Respecting the interests of those who fund, contribute to, use or are affected by statistical practice, and not using statistical practice to mislead any stakeholder.
• Principle D: Responsibilities to Research Subjects, Data Subjects, or Those Directly Affected by Statistical Practices. Adhering to rules for the protection of human and animal subjects and not misusing data.
• Principle E: Responsibilities to Members of Multidisciplinary Teams. Working with colleagues from other professions on a shared understanding of the statistical work.
• Principle F: Responsibilities to Fellow Statistical Practitioners and the Profession. Obligations to peers and to the profession across the wide range of settings in which statistical practice occurs.
• Principle G: Responsibilities of Leaders, Supervisors, and Mentors in Statistical Practice. Specific obligations to follow and promote the guidelines, and to protect practitioners who comply with them.
• Principle H: Responsibilities Regarding Potential Misconduct. Knowing institutional procedures, clarifying facts before alleging misconduct, lodging complaints discreetly with the relevant institutional body, and insisting on a fair process.

Who Do the Guidelines Apply To?

The guidelines apply to all statistical practitioners, whatever their title. This includes academic researchers, data scientists and analysts in industry or government, because the definition rests on the activity, not on a credential. The ASA states that all statistical practitioners are expected to follow the guidelines and to encourage others to do the same. The appendix addresses organizations and institutions that engage in statistical practice, including their duty to protect practitioners who comply with the guidelines.

What Do the Guidelines Say About Data, Methods and Machine Learning?

Because “statistical practice” explicitly includes model or algorithm development and deployment, the guidelines reach machine learning and other predictive work. Under Principle B, practitioners are asked to communicate data sources, known biases and processing or transformation steps, including missing data handling, and to be transparent about assumptions and algorithmic biases. For models and algorithms that inform or implement decisions repeatedly, Principle B calls for plans to validate assumptions and assess performance over time, and to consider criteria and mitigation plans for failure and retirement.

How Do the Guidelines Address P-Hacking and Multiple Comparisons?

The guidelines do not use the word “p-hacking”. Principle B asks practitioners to be transparent about a priori versus post hoc objectives and planned versus unplanned practices, and to disclose when multiple comparisons are conducted and any relevant adjustments. Principle C asks practitioners to differentiate exploratory from confirmatory findings so stakeholders can understand and apply the results. Principle A states that it is contrary to the guidelines to report or follow only results that conform to expectations without explicitly acknowledging competing findings.

How Can Researchers Use the Guidelines?

1. Read the official text. The full guidelines, with a PDF download, are published by the ASA at amstat.org/your-career/ethical-guidelines-for-statistical-practice.
2. Check study design against Principles B and C. Decide in advance which objectives are confirmatory and which are exploratory. Teams comparing methods can review data science vs. statistics or how to choose a statistical test, and the z-test vs. t-test comparison.
3. Check data protections against Principle D. Confirm that applicable rules, approvals and ethical review requirements for human and animal subjects have been identified.
4. Document data and methods. Principle B asks for communication of data sources, processing steps and assumptions, and for sharing the data used as allowable. Publishing data with a persistent identifier such as a DOI can support this.
5. Use Principle E with colleagues from other fields. Use it as the basis for agreeing on what the statistical work can and cannot support.

Strengths and Limits

Strengths

• Defined by activity. The guidelines cover anyone engaged in statistical practice, regardless of job title or field of degree.
• Explicit methodological expectations. Principle B names concrete behaviors such as disclosing multiple comparisons and handling of missing data.
• Shared duty. Principle G and the appendix place responsibilities on leaders, supervisors and organizations, not only on individual analysts.

Limits

• Professional guidance. The guidelines are a statement of professional standards approved by the ASA Board; the document itself describes expectations, not legal penalties.
• Judgment is required. The ASA notes that principles may require balancing competing interests, and that an unexpected ethical challenge calls for seeking guidance in the guidelines.

FAQ

What are the eight principles of the ASA Ethical Guidelines for Statistical Practice?

(A) Professional Integrity and Accountability; (B) Integrity of Data and Methods; (C) Responsibilities to Stakeholders; (D) Responsibilities to Research Subjects, Data Subjects, or Those Directly Affected by Statistical Practices; (E) Responsibilities to Members of Multidisciplinary Teams; (F) Responsibilities to Fellow Statistical Practitioners and the Profession; (G) Responsibilities of Leaders, Supervisors, and Mentors in Statistical Practice; and (H) Responsibilities Regarding Potential Misconduct.

When were the guidelines approved?

The current version was approved by the ASA Board of Directors on February 1, 2022. It was prepared by the Committee on Professional Ethics of the American Statistical Association.

Do the guidelines apply to data scientists and machine learning work?

Yes. The ASA defines statistical practitioners as everyone who engages in statistical practice, regardless of job title, profession, level or field of degree, and defines statistical practice to include model or algorithm development and deployment.

What do the guidelines say about pressure to change results?

Principle A says the ethical practitioner opposes efforts to predetermine or influence the results of statistical practices and resists pressure to selectively interpret data. Principle H covers potential misconduct, including lodging complaints discreetly with the relevant institutional body.

Who is responsible for ethical practice besides individual statisticians?

Principle G sets obligations for those who lead, supervise or mentor statistical practitioners, and the appendix sets responsibilities for organizations and institutions that engage in statistical practice.

Sources

• American Statistical Association. “Ethical Guidelines for Statistical Practice.” Approved February 1, 2022. Checked November 7, 2026. https://www.amstat.org/your-career/ethical-guidelines-for-statistical-practice