Bias-free Language In Research Writing

Scientific and medical articles often include information about human participants. Biased language that categorizes people on the basis of their sex, race or ethnicity, age, disabilities, or sexual orientation can be offensive and inaccurate and will distract your reader. In this blog post, I summarize advice from scientific and medical style guides^1–3  to help you keep your research writing inclusive and bias free.

Your research article should not contain any irrelevant information about the individuals you have studied. Only refer to a participant’s sex, race or ethnicity, age, disabilities, or sexual orientation if it is scientifically or medically relevant to your research question. Take a look at the following examples:

Original: Thirty-four patients, including six African Americans, were eligible for inclusion in the study (contains bias: the ethnicity of the included participants was not relevant to the research question).

Revision: Thirty-four patients were eligible for inclusion in the study (bias free).

Original: All physicians (six of whom were women with children) working regular night shifts in our centre completed the questionnaire (contains bias: the sex of the physicians and whether or not they have children were not relevant to the research question).

Revision: All physicians working regular night shifts in our centre completed the questionnaire (bias free).

Now let’s take a closer look at some topics that commonly attract biased language.

Gender-neutral language should be used wherever possible in research papers. He as a gender-neutral pronoun is a common source of gender bias as it excludes women. For example:

When a patient follows his doctor’s instructions, he is more likely to recover quickly (gender biased).

There are solutions to this problem. You can use a plural noun and pronoun:

When patients follow their doctors’ instructions, they are more likely to recover quickly.

This is better, but not helpful if you are talking about one individual. You can also replace the pronoun with an article:

When a patient follows the doctor’s instructions, that patient is more likely to recover quickly.

But this can lead to awkward phrasing and disrupt the readability of your text. A more elegant solution is the singular they:

When a patient follows their doctor’s instructions, they are more likely to recover quickly.

Some scientific style guides still grumble about this and prefer that agreement in number be maintained in scientific writing, but other style guides have accepted it. If the style guide of your chosen journal does not object to the singular they, use it.

Do not assume the gender of your peer reviewer when responding to their comments on your manuscript. For example:

We thank the reviewer for his insightful comments, which have helped us to improve our manuscript

is likely to annoy your reviewer, especially if she is a woman. An easy solution to this problem (if the singular they is frowned upon by your journal) is to address the reviewer directly:

We thank you for your insightful comments, which have helped us to improve our manuscript. 

Do not specify the sex of physicians, nurses, and scientists (e.g. female physicians or male nurses) unless you want to specify that everyone in the group was the same sex (e.g. we surveyed all female physicians working in our centre between September 2013 and March 2016 is acceptable). 

In clinical studies, it is often necessary to specify how many participants were male and how many were female because the biological differences between the sexes can have implications on the findings. However, use male and female as adjectives rather than nouns to avoid dehumanizing the participants. For example:

Fifty participants were male and twenty were female 

is better than

Fifty males and twenty females were recruited for the study.

Avoid the phrase man and wife, which defines the woman in terms of her relationship with the man. Husband and wife or man and woman is equal and acceptable.

Here are more suggestions for gender-neutral language in your writing:

People, humans, humanity, humankind instead of mankind

Employees, personnel, workforce instead of manpower

Artificial or synthetic fibre instead of man-made fibre

Chairperson or chair instead of chairman or chairwoman (unless talking about a particular man or woman).

Define the racial or ethnic origins of your study participants only if it is scientifically or medically relevant to your research question. For example, some genetic disorders are more common in individuals with specific ancestral origins because the disease-causing mutations are passed down from common ancestors (e.g. Tay–Sachs disease is more common among Ashkenazi Jews). In these cases, it is relevant and acceptable to describe the ethnic origins of your participants.

Explain why race was assessed and justify the inclusion or exclusion of certain groups in your paper. Be as specific as possible when referring to racial and ethnic groups – an individual described as Asian American may have ancestral origins in one of many countries. Use a more specific designation, e.g. Chinese American, Bangladeshi American, or Vietnamese American.

Use commonly accepted designations to define the race or ethnicity of your study participants. Also be sensitive to how your participants identify themselves if possible (do they consider themselves to be Black or African American, Hispanic or Latino?). Avoid referring to non-White racial groups as racial minorities as this can be viewed as being less than the majority (White).

Make sure that the language you use to describe disabled people respects their integrity as human beings. For example, do not use the phrase confined to a wheelchair. Wheelchairs are not prisons – they allow their user to move freely.

Do not suggest that a physical disability is an obstacle to being able to accomplish non-physical tasks. For example:

Although the patient used a wheelchair, she was still able to complete the questionnaire

is not acceptable.

Scientific and medical style guides agree that labelling people with their diseases or disabilities should be avoided. People-first language is preferred when talking about people with a disability or disorder. For example, persons with diabetes is considered more acceptable than diabetics.

But consider the preferences of your participants. Some people strongly feel that their condition is a part of who they are, and can be offended by people-first language. For example, many members of the autism community prefer to be referred to as autistic rather than a person with autism.

Don’t refer to people with diseases as victims, afflicted, maimed, and sufferers as these terms suggest helplessness. For example, use person with cancer not cancer victim. 

In health-related research papers where people with diseases or disabilities are referred to many times, it is acceptable to use the adjectival form (e.g. cancer patient instead of person living with cancer) to improve the readability of the text.

Negative descriptions of conditions should be avoided. For example:

Intellectual disability is better than mental retardation

Substance use disorder is better than substance abuse problem

Brain injury is better than brain damage

Physically disabled is better than crippled or deformed.

Only refer to a person’s sexual orientation if it is scientifically relevant to the data you are presenting. Avoid the term sexual preference as this implies the person can (and maybe should) choose who they are sexually attracted to. The terms lesbians, gay men, bisexual men, and bisexual women are preferred over the more general homosexual as these terms are more specific.

Take care not to confuse case and patient: a case is an occurrence of a disorder or illness, and a patient is a person affected by a disorder or illness who is receiving treatment. A common mistake is to refer to the treatment of a particular case. Take a look at the following example:

Schizophrenic cases were treated with antipsychotic medication (Incorrect)

Patients with schizophrenia were treated with antipsychotic medication (Correct).

Also take care not to refer to the patient when you mean to refer to the treatment. For example:

Patients were managed according to their symptoms (Incorrect)

Treatment was managed according to the patient’s symptoms (Correct).

Use the active voice rather than the passive when describing how individuals participated in your study to avoid giving the impression that they were acted on. For example:

Participants completed the questionnaire

is better than

We administered the questionnaire to participants.

Also avoid using failed to describe an individual’s participation in the study. For example:

Four patients did not meet the eligibility criteria

is better than

Four patients failed to meet the eligibility criteria 

and

Two participants did not complete the questionnaire

is better than

Two participants failed to complete the questionnaire.

Don’t use words and phrases that may be ambiguous in different parts of the world. For example:

We recruited participants in early spring and finished collecting our data by autumn the next year

is not clear to everyone because spring occurs at different times of the year in different parts of the world. Your reader should not have to look up your address to figure out what you mean:

We recruited participants in April 2018 and finished collecting data by October 2019

is more universal.

Bias-free language avoids unnecessary offence and promotes scientific accuracy in research writing. Following the advice outlined in this article will help you to use inclusive language and avoid unnecessary bias in your next research paper.

There is no place for bias and stereotypes in research writing

1. AMA Manual of Style (10th Edition). Oxford University Press.
2. Scientific Style and Format (8th Edition). The University of Chicago Press.
3. Publication Manual of the American Psychological Association (6th Edition). American Psychological Association.

Claire Bacon is a former research scientist with professional qualifications in copyediting and medical editing. She edits scientific research papers and teaches courses on scientific writing.