Race and ethnicity in research

Concept

Race and ethnicity are social constructs; in research they are recorded by self-identification and interpreted with care.

Updated Nov 13, 2026

Key points

Medical journal editors' recommendations ask authors to define how race or ethnicity was determined and to explain its relevance; they treat them as social constructs.
The preferred method is self-identification: the person states which group they belong to; classification by an observer introduces error.
A difference between ethnic groups is a signal of inequity that is usually explained by social determinants and racism, not by genes.
In Colombia, DANE's 2018 census asked about ethnic self-recognition: Indígena; Gitano(a) o Rrom; Raizal del Archipiélago de San Andrés, Providencia y Santa Catalina; Palenquero(a) de San Basilio; Negro(a), mulato(a), afrodescendiente, afrocolombiano(a); or Ningún grupo étnico.
Misclassification (for example, when the question is not asked) dilutes differences and hides inequities; DANE analysed this problem in its 2018 census.
With small groups, report confidence intervals and avoid merging categories without a reason.

What it is

Race and ethnicity are social and historical categories, not biological ones: genetic variation is greater within each group than between groups. Even so, they matter for health because they reflect differences in access, living conditions and discrimination that produce inequities. In research they are collected by self-identification, studies explain how and why they were measured, and they are not used as a stand-in for biology or ancestry.

Worked example

Data (made up): 1 000 pregnant women in a hospital; 120 self-identify as Black or Afro-Colombian and have 6 cases of severe maternal morbidity.
Rates by self-identification: 6 / 120 = 50 per 1 000; the rest, 22 / 880 = 25 per 1 000; RR = 50 / 25 = 2.0.
If an observer records ethnicity, only 80 of the 120 are classified correctly (3 cases); 40 (with 3 cases) move to the rest.
Rates with observer classification: 3 / 80 = 37.5 per 1 000; the rest, 25 / 920 ≈ 27.2 per 1 000.
RR with observer classification: 37.5 / 27.2 ≈ 1.38; the inequity looks much smaller.
Reading: misclassification hides almost two thirds of the excess risk; an RR of 2.0 calls for studying access and quality of care, not biology.

Practices when recording ethnicity in a study

Practice	Recommended	Problem it avoids
Self-identification	Yes, with the official question and categories	Classification by appearance
Explain why it is measured	Yes, in the methods	Using ethnicity with no hypothesis
Use ethnicity as a stand-in for genes	No; if ancestry matters, measure it directly	False biological conclusions
Analyse social determinants	Yes: access, income, territory, discrimination	Blaming the inequity on the group
Language	Neutral and respectful, using the names the groups themselves use	Stigma

Markers

Self-identification
DANE
NARP
Differential approach
Inequity

In practice

In Colombia ethnic groups have constitutional recognition, and data by self-recognition guide policies with a differential approach.
A medical record or survey uses the same question and categories as DANE so the data are comparable.
When interpreting a difference, look for explanations in territory, insurance, access barriers and discrimination.

Common pitfalls

Classifying ethnicity by appearance or surname.
Using race as a biological explanation for a difference without measuring social determinants.
Merging distinct groups into one category to ‘get bigger numbers’.
Using terms the groups themselves do not recognise or that stigmatise.

Test yourself

Think of the answer, then tap the question to see it.

Why is race considered a social construct?

Because its categories change with history and society, and human genetic variation is greater within each group than between groups.

Which method is recommended for recording ethnicity in a study?

Self-identification: each person states which group they belong to, using a defined question and categories, ideally the country's official ones.

What does misclassification of ethnicity usually do to a difference between groups?

It usually dilutes it: by mixing people from different groups, the rates move closer and the inequity looks smaller than it is.

Which categories did DANE's 2018 census use in its ethnic self-recognition question?

Indígena; Gitano(a) o Rrom; Raizal del Archipiélago de San Andrés, Providencia y Santa Catalina; Palenquero(a) de San Basilio; Negro(a), mulato(a), afrodescendiente, afrocolombiano(a); Ningún grupo étnico.

Add to my cards (4)

Connections

← All of epidemiology

Educational content on epidemiological and statistical methods; the examples use made-up numbers to teach the calculation. Not a clinical or legal guide: current Colombian rules are in the official sources.

Sources: Población negra, afrocolombiana, raizal y palenquera. Resultados del Censo Nacional de Población y Vivienda 2018 (Departamento Administrativo Nacional de Estadística (DANE), Colombia, 2019); A. Preparing a Manuscript for Submission to a Medical Journal (International Committee of Medical Journal Editors (ICMJE), 2026); Social determinants of health (World Health Organization, 2026).

Spotted a mistake?We are always improving for the community. Tell us what is wrong and our team will review it.Report a mistake

Representative model, for educational purposes only.