This guide is for students, staff, and faculty who are incorporating an anti-racist lens at all stages of the research life cycle.

Statement on bias in research

Bias is a preference for one thing over another. Research and data are inherently biased due to preferences for certain perspectives. While researchers try to decrease bias in their work it is impossible to eliminate bias completely. Explicit and implicit (unconscious) bias can be found in data sampling, collection methods, analysis and how the data is shared. Bias is present on a systemic level in what gets researched, published, and cited. As you are researching your topic ask yourself: What perspectives are being favored? What perspectives are missing? Why? How can I add missing perspectives?

De-center whiteness in primary research

Oftentimes we are taught about the distrust of Black, Indigenous, and People of Color (BIPOC) due to research atrocities (e.g. Tuskegee Study, forced sterilization), but this just places the blame on those communities for not participating in research (i.e., “there’s nothing we can do because they don’t trust us”). This distracts from the fact that researchers have left communities of color out of research due to white centering. The strategies below are designed to de-center whiteness, think inclusively, and build trust between researcher and communities of color when conducting primary research.

 

Solution: Evaluate whether your research is WEIRD

Henrich, Heine, and Norenzayan (link to article below) propose that most published research is not representative of the majority of populations because it was conducted with WEIRD societies.

  • Western
  • Educated
  • Industrialized
  • Rich
  • Democratic

When developing a research design, ask yourself how you can de-center the status quo characteristics described by WEIRD. SOPHIA TO ADD CONTENT HERE.

 

Solution: Recruit BIPOC people and communities for inclusion in studies

Funding agencies are implementing policies that require diversity in study recruitment (e.g., National Institutes of Health Inclusion of Women and Minorities as Participants in Research Involving Human Subjects), and this is a particular concern for clinical trials. The following are strategies for de-centering whiteness in clinical trial recruitment (note that the use of the term "minorities" is still white-centering):

 

Solution: Utilize research methods and practices that de-center whiteness

A Toolkit for Centering Racial Equity Throughout Data Integration, by Actionable Intelligence for Social Policy, helps researchers embed questions of racial equity throughout the data life cycle: planning, data collection, data access, algorithms/use of statistical tools, data analysis, and reporting and dissemination. It includes exercises and examples. See the excerpt below:

 

For more information on de-centering whiteness in primary research, see:

De-center whiteness in secondary research

Secondary research involves the summary, collation, and/or synthesis of existing data or research, and often involves deep dives into existing literature. The strategies below are designed to de-center whiteness, think inclusively, and build trust between researcher and communities of color when conducting secondary research.

 

Solution: Evaluate whether your research is WEIRD

Henrich, Heine, and Norenzayan (link to article below) propose that most published research is not representative of the majority of populations because it was conducted with WEIRD societies.

  • Western
  • Educated
  • Industrialized
  • Rich
  • Democratic

When developing a literature search, ask yourself how you can de-center the status quo characteristics described by WEIRD. SOPHIA TO ADD CONTENT HERE.

 

Solution: Follow the praxis put forth by the Cite Black Women Collective

  1. Read Black women's work
  2. Integrate Black women into the CORE of your syllabus (in life & in the classroom)
  3. Acknowledge Black women's intellectual production 
  4. Make space for Black women to speak 
  5. ​Give Black women the space and time to breathe

These principles from the Cite Black Women Collective can be applied to all BIPOC.

 

Solution: Consult research found in non-Western journals

The Journals Online Project - aimed at providing increased visibility, accessibility, and quality of peer-reviewed journals published in developing countries so that the research outputs produced in these countries can be found, shared, and used more effectively - was launched by the International Network for Advancing Science and Policy (INASP) in response to voices not heard, wasted talent, and unused research.

 

For more information on de-centering whiteness in secondary research, see:

Acknowledge that scholarly publishing is racist

Peer review for is meant to ensure rigid methodology and low bias in what gets published, but that system is flawed. Most editors and peer reviewers are white, making implicit bias a core problem, and research on race and racism are not favored by gatekeepers. This problem also acts to exclude scholars who are Black, Indigenous, and People of Color (BIPOC) and/or from low-income countries.

 

Solution: Look outside peer reviewed literature for perspectives from marginalized voices and groups

Researchers should look at the gray literature for perspectives from marginalized communities. Gray literature refers to works published outside traditional methods, and the easiest way to access it is through Google. There is no perfect solution, and many strategies will be attempted before useful information is found. Google search tips:

  • Keep in mind that Google itself uses biased algorithms, so searching must be specific and incorporate various terms to represent one concept.
  • The following example shows one possible search strategy for housing discrimination: "marginalized voices in housing."
  • You might also try using keywords of interest plus "site:org" which tells Google that you are looking for organization rather than a company or government site.

Archives also feature marginalized voices through collections of oral histories, documents, and interviews. Where to start:

 

Solution: Search for BIPOC scholars through professional organizations

It is not possible to search for an author's race in literature databases - the following is a flawed, yet workable, solution to that problem: 

  • Google organizations that represent marginalized groups and search their websites for special interest groups, experts, publications, data, etc. For example, perusing the digital magazine available on the National Society of Black Engineer's website to find leaders in the field.
  • Seek out a networking organization, such as ColorComm, which is an essential organization for women of color in all areas of communications including public relations, advertising print media, broadcast, and more. To look for such groups, use a variety of terms - one might be "research organizations run by people of color."
  • SOPHIA TO ADD TO THIS PARAGRAPH.

 

Solution: Utilize smaller, lesser known databases that - by nature - feature scholars who are BIPOC

The list below is not exhaustive, but it features a handful of indexes focused on BIPOC issues as well as a selection of local BIPOC-led news outlets:

 

For more information on bias in scholarly publication, see: 

Acknowledge that search algorithms are racist

Algorithms existed pre-internet, but in today's world, they serve as a sequence of instructions to perform computation. There is evidence that algorithms - even those in academic databases - are sometimes racist. Proprietary algorithms (e.g., Google) are customized to users and lack transparency, making it unclear why search results vary from person to person. For this reason alone, you may be missing important information unless you know to search specifically for it.

 

Solution: Use inclusive search terminology on topics of racism

Terminology used to describe race and ethnicity have evolved over time. This variability in language can make searching comprehensively for literature on race/ethnicity difficult. If you are performing a comprehensive literature search (e.g., systematic review, scoping review, historical perspective), you must include outdated terminology in your search strategy or check to see if old terminology has been included in the databases’ subject headings. And other times, you will need to use terms that were common from the preferred date range of the research.

This health sciences example from Ovid MEDLINE shows that older terminology from 1963 forward has been added to the new subject heading “African Americans.”

Searching for literature about racism requires a sophisticated search strategy to not only include “racism,” but also biases and discrimination against specific races and ethnicities. The example below is a search strategy for racism in Ovid MEDLINE. It uses both MeSH terms and keywords for racism, and also MeSH terms and keywords for bias, discrimination, stigma, stereotyping, and prejudice grouped with MeSH terms and keywords for specific races and ethnicities. Remember that terms evolve. For example, BIPOC (Black, Indigenous, and People of Color) and BBIPOC (Black, Brown, Indigenous, and People of Color) should be added to the racism search example below, which was only constructed a year ago.

 

For more information on bias in algorithms, see: 

Acknowledge that library cataloging systems are racist

When incorporating anti-racism into research, it’s important to acknowledge the context in which information has been shared through library systems. Dewey Decimal, the Library of Congress, and smaller discipline-specific cataloging approaches were designed in a racist and white-centered system. The following excerpt from The Racist Problem with Library Subject Classifications (link to article below) provides an example of the challenges this creates for a linguist researcher, for example: "Of the twelve subclasses dedicated to languages, seven are dedicated to European (western) languages. Non-western languages do not get this same courtesy. 'Oriental languages,' ...Indo-Iranian languages, all languages from Eastern Asia, Africa, and Oceania, and Hyperborean, Indian, and 'artificial' languages are each lumped together into four classifications (PJ, PK, PL, and PM, respectively). Not only is it racist to lump all non-western languages, usually spoken by POCs, into generalized categories ...but it’s also incredibly problematic that the term 'artificial languages' is categorized with non-western languages, othering them and implying that the non-western languages are not real and should not be taken seriously." 

 

Solution: Know that librarians are fighting for change

As a researcher, it is important to be aware of this information, but there is no action you can take to change it. Just know that librarians are advocating for anti-racist cataloging, a long-term process.

 

For more information on bias in cataloging, see: 

Who we are

This guide was developed by two librarians and a library intern from the University of Minnesota (UMN) - all white-identifying - with invited input from all other UMN librarians. We approached this project from the perspective of library staff at a large research institution. Ours is not the only perspective - perhaps we are not seeing the ways that this guide misses or marginalizes other points of view. We invite you to contact us to share your input for consideration for website content (Click on the "Email contact form" link on the left panel).

Last Updated: Dec 1, 2020 3:17 PM