Recommendation letters and discrimination

A study conducted by researchers at Rice University examined reference letters written for academics in support of job applications or job promotions. The study examined whether such letters reflected gender stereotypes, and if so, what the impact might be on job and promotion prospects. Unsurprisingly, they discovered that letter writers did employ gender schemas, and these were likely to have a negative impact on women’s careers. The researchers reviewed 624 letters of recommendation, written for 194 people who applied for junior positions at a US university. Letters of recommendation written for women tended to describe them in more ‘communal’ (social, emotive) terms, using words such as ‘affectionate’, ‘helpful’, ‘caring’, ‘sympathetic’ and ‘tactful’. Men were described in ‘agentic’ (assertive, active) terms, such as ‘confident’, ‘forceful’, ‘independent’, and ‘daring’. The researchers then anonymised all the letters, controlled for awards, publications, experience, and so on, and gave the letters to faculty to evaluate. The letters that described the applicant in more ‘communal’ terms consistently received more negative evaluations. In addition, letter writers tended to be more hesitant about women’s qualities – e.g., letter writers for women were more likely to say things like ‘I think she will probably make a good leader’, whilst letter writers for men tended to say things such as ‘He is an established leader’. If this phenomenon is widespread, and there is reason to suppose it is, then it amounts to a form of implicit discrimination – whilst there is no suggestion that letter writers or readers are consciously employing gender schemas/rating women more negatively, the net result is to disadvantage women applicants. A second study is planned, which will examine the same issues with respect to medical jobs. You can read a write-up of the study here.

12 thoughts on “Recommendation letters and discrimination

  1. I do not understand why letters of recommendation are so important. Don’t they say more about the qualities of your network, rather than the qualities of your work? It is theoretically possible that someone who is weak as a philosopher would nevertheless get glowing letters of recommendation (such as ‘He is among the top 2 % of my students of the past 20 years) if they have a good network (i.e., went to fancy grad school, do buddy-buddy activities with the supervisor etc.) My personal experience tells me that this kind of networking is still something that mainly males are good at (for one thing, at least in my department, all professors are male, so it is also more easy for males to network.)

  2. I think letters of recommendation are important (I’m describing the fact, not necessarily endorsing it) – say someone applies for promotion and their case is assessed by a committee of non-experts in their field, quite likely – in the UK at least – no philosopher on it at all. What else can they go on other than recommendation letters?
    A related comment. My suspicion, based on the UK and talking to various colleagues and individuals, is that there is a level of widespread implicit bias against women at the stage of applying for promotion. In the UK this woks by the individual having to get their head of department to support an application. Anecdotal evidence from quite a lot of people suggests to me that heads are more quick to support applications from men. With women, there is more often a reaction of ‘leave it for a couple of years’, ‘you might not be ready yet, leave it a bit longer to be sure’, whereas with men it’s more ‘yes let’s give it a try’. I don’t know what the underlying schemas are here, but would be very interested to hear of any research that’s been done into this or any ideas on what the underlying schemata might be. I would also guess that if a head hesitates, and a man asserts he wants to go ahead anyway, the head would be more likely to react favourability to that than if a woman asserted her case in the same situation.

  3. What I found particularly interesting from another study is that letters of rec for women have a signiicantly higher significant percentage of doubt raisers (25 % of letters for women have doubt raisers versus 12 % for men, see Doubt raisers are phrases like ‘while not the best student I ever had’, or ‘despite her multiple pregnancies, she still had a satisfactory output’, or ‘in spite of her personal struggles…’

  4. I found a copy of my letters lying around on a table after an APA reception and, indeed, the very phrase “in spite of her personal struggles” came out–with a narrative about my husband’s illness with which I was dealing while in grad school.

    I’ve seen it from the other end too, in letters for candidates I read, and at department meetings. And in my experience in hiring has been:

    X’s research record and potential is terrific, but we won’t make her an offer because she’ll get a zillion offers because of affirmative action since she’s a woman.

    Y is every so nice so we’ll fly her in. Jeez, Y is a dud: why did we waste the money?

  5. I recognize the previous commentor’s experience well. In my department, they have been thinking about hiring a woman, since all members are male (and it’s a large department). They are not considering to promote any of the talented women philosophers (postdocs and ABDs) here because “they would find a job pretty easily anyway, what with all the affirmative action going on”, but in the meantime they are wondering “how can we attract a good female philosopher”?

  6. I wonder whether this study gets to the heart of the problem. The study seems to suggest that implicit discrimination (or perhaps better, gender stereotyping) on behalf of letter writers causes women to be framed in ways that are hurtful to their job prospects. For that claim to be true, the conditional “if female applicants were not framed in these ways by letter writers, their prospects would be better” must itself be true. But I wonder if it is true. Letter writers are not the only ones who perceive the world through gender stereotypes: so do search committee members. And they may evaluate female applications worse (or no better) if they are framed in gender neutral ways. Men are assertive, women are bitchy; men are confident; women are pushy (that is, the same behaviors are perceived differently by observers as a consequence of their stereotypes). So saying, eg., “she is an established leader” may not cause a female applicant to be perceived more positively than saying “she is a caring individual”.

  7. Thanks so much for posting this study. I found myself sympathetic with H, the original commenter. I think the more I talk to folks about the LoRs for job hires, the less I’m convinced they do much useful work. I can certainly see how negative letters would serve to rule out some candidates, but I fail to see how the letters are going to make much headway in separating out pretty good/good/great/excellent candidates. It reminds me of quantitative student evals when used to evaluate performance, quite effective at separating the horrendous from everyone else, but not useful for anything beyond. I’ve listened (I’m a grad student, so no first-person, direct participation) to faculty talk about some of the finer points of the ‘code’ used in these letters, and I’m rather not convinced that it does any useful work.

    And the doubt raising statistics are very troubling.

  8. That’s certainly a possibility, Neil. I guess another good study that could be carried out could test for that by asking people to evaluate the same female applicant, but with different styles of letter to see whether the different letter results in different assessment.

Comments are closed.