Avoiding Unintended Gender Bias in Letters of Recommendation (Case Study 1)

Reducing Unconscious Bias to Increase Women’s Success in IT

In an analysis of 300 letters of recommendation for research and clinical faculty positions at a medical school, researchers concluded that recommenders often unconsciously describe candidates in stereotypically gendered ways. Trix and Psenka (2003) found that compared with letters written on behalf of men, letters written about women were shorter and more likely to lack basic features, such as how they knew the applicant, concrete references about the applicant’s record, or evaluative comments about the applicant’s traits or accomplishments. The researchers also found that descriptions of men were more likely than those of women to be aligned with the critical job requirements: research record and ability.

NUANCED, HIDDEN BIASES IN LANGUAGE

Stereotypical and Grindstone Adjectives. Adjectives used to describe both male and female applicants were often based in gender stereotypes: men as successful and women as nurturing. Words like “compassionate” were frequently used for women, while words like “accomplishment” were more often used for men. Grindstone words — adjectives describing applicants as hard workers — were also more often used for women than for men, implying that women may have strong work ethic, but men have ability.

Repetition of Standout Words. Word and phrase repetition leads to cohesion and can be a persuasive rhetorical device. When superlatives and status words (e.g., “outstanding,” “research”) were used in letters, they were repeated more often in letters describing men than women.

Doubt Raisers. Letters recommending women were twice as likely to include “doubt raisers” (e.g., “it appears that her health is stable”; “while she has not done…”) than letters written for men.

Men Research, Women Teach. Letters describing the positive qualities of men more often emphasized their role as researchers and professionals, while letters describing the positive qualities of women more often emphasized teaching. The pronoun “her” was followed by “training,” “teaching,” or “personal life” much more often than was “his.” Similarly, “his” was more often followed by “research,” “skills,” and “publications” than was “her.”

REDUCING UNCONSCIOUS BIAS IN LETTERS OF RECOMMENDATION

Consider the following when checking letters you write for bias.

  • Focus on comparing the applicant with the requirements of the job.
  • When describing stereotypically female traits, ask yourself if these characteristics are relevant to the job and if you are missing other strengths.
  • Avoid overuse of gendered or grindstone adjectives.
  • Avoid unnecessarily invoking a stereotype (“she is not emotional…”).
  • Use title and surnames for both men and women instead of first names, unless using first name is standard in your field.
  • While it is usually important to talk about the personality and interpersonal skills of the applicant, avoid overly focusing on them.

The example below is from a pair of real letters of recommendation written for a job candidate before and after the writer learned about unconscious bias in letters. The “before” example needlessly repeated stereotypical expectations for women (she was nice, hardworking, and easy to get along with). The “after” example was revised to focus more on the requirements of the postdoctoral positions for which she was applying.

REVISING FOR FOCUS ON REQUIREMENTS

Overly focused on interpersonal skills: a gender stereotype

…quite gifted interpersonally: she is easy to get along with and quick to understand social situations.

…cares about her work and the needs of others around her.

Focused on the technical requirements of the position

…with the necessary scientific methodological expertise to result in a complex and insightful dissertation study.

…technically skilled, deeply knowledgeable, resourceful, success oriented, and a pleasure to work with.



References

  • Trix, F. & Psenka, C. (2003). “Exploring the color of glass: Letters of recommendation for female and male medical faculty.” Discourse & Society, 14(2), 191-220.

View related research:

View related case study:


Authors: Lecia Barker