Ethical Guidelines for Statistical Practice

American Statistical Association
Committee on Professional Ethics

I. Preamble
A. Purpose
The ASA Ethical Guidelines for Statistical Practice are intended to help statistical practitioners make and communicate ethical decisions. All professional users of statistical methods should urge clients, employers, researchers, policy makers, journalists, and the public to expect statistical practice to be in accordance with these guidelines and to object when that is not the case. Students learning how to apply statistical theory to problems should be encouraged to follow these guidelines whether or not their target professional specialty will be "statistician."

B. Statistics and Society
The professional conduct of statistical analyses is essential to many aspects of society. The use of statistics in medical diagnoses and biomedical research may affect whether individuals live or die, whether their health is protected or jeopardized, and whether medical science advances or gets sidetracked. Life, death, and health, as well as efficiency, may be at stake in statistical analyses of transportation, occupational, or environmental safety. Early detection and control of new or recurrent infectious diseases depends on sound epidemiological statistics. Mental and social health may be at stake in psychological and sociological applications of statistical analysis.
Effective functioning of the economy depends on the availability of reliable, timely, and properly interpreted econometric data. The profitability of individual firms depends in part on their quality control and their market research, both of which should rely on statistical methods. Agricultural productivity benefits greatly from statistical applications to research and to output reporting. Governmental policy decisions regarding public health, criminal justice, social equity, education, the environment, siting critical facilities, and other matters depend in part on sound statistics.

Scientific and engineering research in all disciplines require careful design and analysis of experiments and observations. To the extent that uncertainty and measurement error are involved; that is, in most research; experimental design and analysis are crucially dependent on statistical methods. Even in theory, much of science and engineering inherently involves statistical variability.

C. Shared Values
Because of the dependence of society on sound statistical practice, all statistical practitioners have social obligations to perform their work in a professional, competent, and ethical manner. This document is directed to those whose primary occupation is statistics. Still, the principles expressed here should also guide the statistical work of professionals in all other disciplines which use statistical methods. All statistical practitioners are obliged to conduct their professional activities with responsible attention to:

1. The social value of one’s work and the consequences of how well or poorly it is performed.

2. Pressures or temptations to slant statistical work toward predetermined outcomes. It is all right to advocate a position; it is not all right to misapply statistical methods to gain (only) an appearance of support for that position.

3. Statistics as a science. As in all science, understanding evolves. Statisticians have a body of established knowledge, but also many open issues which deserve frank discussion.

4. Maintaining and upgrading competence in one’s work.

5. Collegiality and civility with fellow professionals.

6. Avoidance of discrimination or harassment relative to: Race, Ethnicity, Sex, Sexual Orientation, Nationality, Age, Religion, or any Disability.

While the following are not ethical obligations so much as good professional citizenship, practitioners who are willing and able to do so may choose to work toward:

7. Improved public understanding of and respect for statistics.

8. Support for sound statistical practice when it is unfairly criticized.

9. Exposure of dishonest or incompetent uses of statistics.

10. Service to the profession as a statistical editor, reviewer, or association official.

II. Ethical Guidelines
A. Professionalism
1. Avoid data selection processes that are inconsistent with transparent treatment of the issues being studied.

2. Use only statistical methodology suitable to the data and to valid results.

3. Strive for valid practical significance, not just statistical significance.

4. Recognize that automated statistical computation alone does not constitute adequate statistical analysis; it is also necessary to understand the theory, the data, and the methods used in each statistical study.

5. Respect and acknowledge the contributions and the intellectual property of others.

6. Disclose conflicts of interest and resolve them through: divestiture of the conflicting personal interest or refusal or withdrawal from the professional activity. Examples where conflict of interest may be problematic include grant reviews, other peer reviews, and tensions between scholarship and personal or family financial interests.

7. Provide only such expert testimony as you would be proud to have peer reviewed.

B. Responsibilities to Funders, Clients, Employers:
1. Offer an analysis of alternative statistical approaches so the best approach can be taken.

2. Clearly state one’s statistical qualifications and experience relevant to one’s work.

3. Clarify the respective roles of different participants in studies to be undertaken.

4. Explain any expected adverse consequences of failure to follow through on an agreed sampling or analytic plan.

5. Apply statistical sampling and analysis procedures without regard for outcome.

6. Keep all statistical methods in the public domain; they are not proprietary, although specific implementations of them may be proprietary.

7. Guard privileged information of the employer/client/funder.

8. Fulfill all commitments.

9. Accept full responsibility for one’s professional performance.

C. Reporting Responsibilities:
1. Maintain personal responsibility for all work bearing one’s name; avoid undertaking work or coauthoring publications for which one would not want to acknowledge responsibility.

2. Report statistical and substantive assumptions made in the study.

3. In co-authored publications, clearly identify the responsibility(ies) for statistical work which affects interpretation of the results and conclusions.

4. The ethically preferred rule for authorship order in statistical publications is by degree of intellectual contribution to the study and to the material to be published, to the extent such ordering can feasibly be determined. Where some other rule of authorship order is used in a statistical publication, the rule used should be disclosed in a footnote or endnote. The primary intent of this rule is to protect the career opportunities of those who need and deserve, but do not always receive, first authorship.

5. Account for all data considered in a study and explain the sample(s) actually used.

6. Report the sources and assessed adequacy of the data.

7. Report the data cleaning and screening procedures used, including any imputation.

8. Report the analytic methodology and its relation to the assumptions.

9. When reporting analyses of volunteer data or other data not representative of a defined population, include appropriate disclaimers.

10. Report the limits of statistical inference of the study and possible sources of error, both random and systematic.

11. Share (non-proprietary) data used in published studies to aid peer review and replication.

12. As appropriate, publicly correct any errors discovered after publication.

D. Responsibilities to Research Subjects:
1. Know and adhere to appropriate guidelines for human subjects protection.

2. Avoid excessive risk or imposition on the time and privacy of research subjects.

3. Protect the privacy and confidentiality of research subjects and the data they provide.

4. Obtain informed consent and fulfill any commitments made in obtaining it.

5. Avoid or minimize the use of deception.

6. Protect special populations appropriately: prisoners, children, pregnant women, mentally or physically handicapped persons, the frail elderly, and others who are not fully and freely capable of protecting their own interests.

7. Know and adhere to appropriate animal welfare guidelines in research involving animals.

E. Responsibilities to Research Team Colleagues:
1. Inform colleagues of other disciplines about relevant aspects of statistical ethics.

2. Promote effective and efficient use of statistics by the research team.

3. Respect the ethical obligations of members of other disciplines as well as one’s own.

4. Assure professional quality reporting of the statistical design and analysis.

5. Avoid compromising statistical validity for expediency.

F. Responsibilities to Other Statisticians or Statistical Practitioners:
1. Promote review of one’s work by sharing data and methods adequately.

2. Be willing to help strengthen the work of others through appropriate peer review. When doing so, complete the review promptly and well.

3. Assess methods, not individuals.

4. Respect differences of opinion.

G. Responsibilities Regarding Allegations of Misconduct:
1. Avoid condoning or appearing to condone incompetent or unethical conduct of statistical studies in one’s working environment.

2. Recognize that differences of opinion and honest error do not constitute misconduct; they may warrant discussion but not accusation.

3. If involved in a misconduct investigation, know and follow prescribed procedures; maintain confidentiality.

4. Following a misconduct investigation, support efforts of the accused, the witnesses, and whistleblowers to resume their careers in as normal a manner as possible.

5. Do not condone retaliation against, or blackballing of, responsible whistleblowers.