Changing pattern of use of ‘ethnicity’ and ‘race’ in scientific literature

R Afshari and R S Bhopal

Department of Community Health Sciences, Section of Public Health Sciences, University of Edinburgh, Edinburgh, UK.

Sir—The International Journal of Epidemiology1 has raised an important question on the shifting preferences in the use of ‘race’ and ‘ethnicity’ in scientific literature. The importance of recognizing these terms as two distinct concepts has been stressed.2 Chaturvedi argued that ‘ethnicity’ is scientifically preferable to race. As the genetic variation between ‘racial’ groups is small, categorization based on biological criteria is inaccurate and misleading in indicating variation in health and disease. Instead, a complex construct such as ethnicity—which surprisingly is still not used in MedLine MeSH heading3—is scientifically preferable.2,4 Studying ethnic variations in health, although socially sensitive, may help to determine aetiology, tackle inequalities, assess need, make public health plans and direct resource allocation.1,4 The lack of consistency in terminology and poor understanding of the concepts may, however, hamper progress in this field and make international collaboration more difficult.4 Ten years ago Sheldon and Parker reported that the use of ‘ethnic groups’ and ‘race’ as variables in the health literature from 1985 to 1990 had markedly increased but that studies on racism were rare.5 We have updated and expanded Sheldon and Parker’s observations.

Search terms, as shown in the Table, were studied in PubMed listed articles and searched in all fields, publication types and languages, during 1966–2000. Based on publication dates, findings were categorized in 5-year periods. The number of articles in each category between 1996–2000 were divided by the number in 1966–1970 to yield a publication growth rate. The search terms were ‘Ethnic groups’, ‘Ethnicity’, ‘Race’ and ‘Racism’. As the latter two terms are not MeSH headings, their MeSH equivalents ‘Racial stocks’ and ‘Prejudice’ were also studied. Some variables unrelated to ethnicity and race (Heart, Depression, Social, and Income) were selected as a reference to the general growth in publication.

Table 1Go shows that while publication relating to all the search terms increased, ethnicity yielded 36.6 times more articles in 1996–2000, compared with 1966–1970, confirming the increasing scientific interest in this concept. In recent years, the number of appearances of ‘Ethnic groups‘, ‘Ethnicity‘, ‘Race’ and even ‘Racial stocks’ are still considerably higher than ‘Racism’ and ‘Prejudice‘, but the gap (in relative terms) has declined, which probably shows the increasing interest in racism. For the four reference search terms (Heart, Depression, Social, and Income) the growth ratios ranged from 2.4 times (for Heart) to 9.9 (for Income). Assuming this range as the background pattern, articles relevant to the MeSH term racial stocks have decreased disproportionately.


View this table:
[in this window]
[in a new window]
 
Table 1 Total number of appearances of search terms of interest in PUBMED listed published articles 1966–2000
 
This analysis shows that ‘Ethnicity’ has been gradually replacing the scientifically and conceptually limited term of ‘Race’ in the scientific literature—a change that will be widely welcomed by most scholars1,2,4,5 Ethnicity should be a MeSH term.

References

1 Chaturvedi N. Ethnicity as an epidemiological determinant—crudely racist or crucially important? Int J Epidemiol 2001;30:925–27.[Free Full Text]

2 Senior PA, Bhopal R. Ethnicity as a variable in epidemiological research. BMJ 1994;309:327–30.[Free Full Text]

3 Entrenz—PubMed. United states national library of medicine; NCBI | NLM | NIH; Department of Health & Human Services; Freedom of Information Act | Disclaimer; http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?CMD=&DB=PubMed, (28–29, March, 2002).

4 Bhopal R. Ethnicity and Race as Epidemiological Variables. London: Taylor and Francis, 2001, pp. 21–40.

5 Sheldon TA, Parker H. Race and ethnicity in health research. J Public Health Med 1992;14:104–10.[Abstract]