Action indeed is the sole medium of expression in ethics.

I give context to the ethical issues surrounding letters of recommendation by summarizing three breaches of ethics by students:

1. A graduate student, suspecting an unflattering evaluation by an advisor based on another grad student’s tip, successfully hacks into the advisor’s computer to read a copy of the letter. The advisor is never aware of the student’s act, and the letter written on the student’s behalf does turn out to be negative.
2. A graduate student, after being denied access to a confidential credentials file, breaks into the graduate office to Xerox the file. The university police become involved and the student is discovered, then kicked out of the program.
3. An undergraduate collects the requested letter of reference in a signed, sealed envelope and becomes nervous about its potential contents. Before sending the letter off to the target graduate program, the student tears open the envelope and reads the favorable letter, then—relieved—sheepishly places the letter in a new sealed envelope with a forged signature across the back.

All these are true accounts that I was privy to in my 20 years of university teaching, and they underscore just how desperate some students become about letters of recommendation, and how some will even take foolish risks just to find out what we’ve written about them.

I have encountered many more typical scenarios, though, where both student and letter writer are uplifted by the process:

1. A graduate student writes a letter of recommendation for the first time, and finds that a 30-minute interview with the student improves the detail of the letter and helps the student win a national scholarship.
2. A faculty member writes a graduating senior a glowing letter that also includes one paragraph of criticism, and discusses the criticism with him as a way to point the student towards self-improvement.
3. An alumna keeps her mentor informed once a year on her progress even six years after her graduation, and thus her mentor is able to write several informed, detailed reference letters for her as new opportunities arise.

These positive examples and others tell us that recommendation letters are not mere formalities involving “paying back favors” we once received from others, nor are they simply redundant paperwork we complete to help students advance—letters of recommendation offer us lessons about relationships (or their lack), growth, power and empowerment, professionalism, attitude, protocol, communication, ethos, and trust. To understand them fully, then, we must consider that the process and act of writing recommendation letters can have a powerful ethical component. This chapter is devoted to fleshing out the ethical issues related to recommendation letter writing, and offering proven strategies on how to address them.

To begin a discussion of the ethical context for writing recommendation letters, we should agree that—at least from the perspective of the person for whom the letter is written—letters can do great good or they can do great harm. Although some faculty question the importance of letters and even speculate as to whether or not they are ever read, others fiercely defend (or attack) their use and relevance, and insist that people’s lives are changed due to letters of recommendation.

Instructive in this regard is a 2002 letter to the editor of The Chronicle of Higher Education by the Chair of the Department of Mathematics and Physics at Troy State University, in which the writer quips: “Scoundrels always seem able to get good letters.” The writer summarizes how he spent a year after completing his PhD staring at the ceiling and getting no interviews, even though he had been assured by a professor that all three recommendation letters in his file were positive. “Only during this period did I come to understand that the aforementioned professor had plotted to keep me from getting a job . . . After 18 months, I resolved not to . . . apply in academia again. I worked in industry, and it was merely by chance that I later came to get full-time work in academia”(1).

In other articles on the subject of recommendation letters, Chronicle readers will find further complaints about foes carrying out vendettas, deliberate obfuscations, parallels between inflated grades and inflationary rhetoric in letters—even calls to abandon the system entirely and pay outside reviewers. On the other hand, readers will also find strong defenses of the current system, acknowledgments that inflationary rhetoric exists but that letters are critical nonetheless, arguments that letters are only one variable among many in the evaluative process, and insistence that letters teach us more about candidates than any other part of their application.

While many of the above examples go to the issue of letters for faculty seeking tenure and promotion, they also illuminate the ethical issues involved as we write letters for our students, who often approach the process of soliciting our aid in something of a nervous haze, not fully aware that none of us achieved our positions without the help of former faculty mentors writing letters of support for us. In a string of favors exponentially repaid, most of us write at least 10 times more (even 100 times more) recommendation letters than we actually received for ourselves; thus we contribute to a system that is only as good as the work we deliver to it. Only by better understanding the system can we hope to improve it.

And there’s nothing new under the sun. Just as modern studies do, studies on letters of reference from the 1920s and 30s show a questioning of the very functions of the documents, concern with the clarity, specificity, and credibility of qualitative praise, arguments about the effect of confidentiality in letters, and open attempts to warn selectors against particular candidates. Consider this excerpt of findings cited in a 1935 study (2), just as relevant today:

1. The writer of testimonials and letters of recommendation is likely to view his task lightly.
2. The writer for mere accommodation will often exceed his knowledge or falsify it.
3. There is no way of checking against errors.
4. Bias or carelessness of the writer is a factor.
5. The writer may overstate or underestimate the case of the candidate.
6. The writer may simply make inadequate statements perfunctory in character.

Interestingly, these same kinds of problems are just as relevant today, suggesting that very little has changed. In deep contrast, recommendation letter writers of old enjoyed far greater candor than they do today. Witness these now ironic excerpts from a 1936 study (3), quoting actual letters written to “recommend” public school teachers:

“Some people in this section have questioned her deportment on certain occasions. . . . I feel that she might do better work in another community.”

“Miss N came to us a year ago. She has been in three different systems in the four years of her experience. . . . We don’t feel that we should prevent Miss N from continuing her annual change.”

“His pupils are fairly well interested in their work, but never excel. I believe you could procure his services at his present salary.”

“She is married but her husband is not with her. . . . If she were not my sister I would like to speak of her in detail.”

“Please destroy this letter when you have read it.”

This pithy last statement is my favorite, in that its request was obviously not carried out. And imagine today if writers commented on someone’s salary or negatively on someone’s marital status in a recommendation letter—certainly a lot has changed in relation to the boundaries of writer commentary.

Considering, then, the substantial power that recommendation letters have to either help or harm the student, and assuming that by the very act of agreeing to write a letter we mean to help, let us begin by recognizing the ethical context in which we write, from the built-in implications to the nuances that we control.

Faithful readers of The Chronicle of Higher Education will realize that academia is increasingly becoming a litigious arena. Faculty members sue their schools over tenure denials; students sue universities over a flap about objectionable website photos; universities sue students for blogging unfavorable comments about administrators. Two of my favorite cases were in 1994, when a University of Idaho student tried to sue after crashing through a third-story window while mooning his friends (4), and two Pace University students sued because they found a class they had enrolled in to be too hard (5). The badly injured mooner sought nearly half a million dollars per cheek (but the judge turned his back), while the Pace University students initially won $1,000 apiece plus couerse tuition reimbursement, but a New York appeals court later overturned the decision.

More to the point, in one of the few instances of legal action over evaluative letters, in a case involving the University of Missouri Medical School, the US Supreme Court upheld that a fourth-year medical student could be dismissed from school as a result of written evaluations criticizing the student’s erratic attendance at clinical sessions, poor interpersonal skills, and lack of “a critical concern for personal hygiene” (6). However, in another case, the University of Pennsylvania failed at repeated attempts to retain sole access to the complete tenure file documents of six faculty, which were requested in a claim of sex and race discrimination (7). In this case, then, letter confidentiality was not supported. Several other legal cases involving faculty letters emerged in the 1990s and 2000s, and the initial rulings in these cases were sometimes overturned on appeal.

Though these legal cases involved faculty and grad students, most letter writers have some sense of worry—often a vague one—of potential legal issues involved in writing letters for undergrads as well. Though actions taken against letter writers are rare, many faculty write letters with a nagging, even if unfounded, fear of legal action, and it certainly influences what they write. In fact, one study of 150 faculty at 150 schools found that “more than half agreed that letters frequently are inflated because writers fear legal retribution for negative comments” (8).

To sort through this problem and consider the relevant issues squarely, we can examine how confidentiality impacts letters and explore the legal interpretations of FERPA (Family Educational Rights and Privacy Act).

The Influence of Confidentiality on Letter Writing

Not surprisingly, studies find that confidential letters contain comments about students that are less favorable than those in open access letters. However, reviewers of letters are also more likely to trust the information when they know that students have waived their access rights (10, 11). In plain terms, everyone is more comfortable when a student opts for a confidential letter, and such a letter will also likely be perceived as more trustworthy. A 2001 article in the journal Academic Emergency Medicine put the matter succinctly: “Knowledge of a candidate’s potential viewing may bias the candid and honest authorship of an accurate letter” (12). Thus, we typically urge students to waive their access rights, and we can do so confidently by assuring them that this is the standard practice held in much higher favor by schools and selection committees.

FERPA’s Impact on Student Records

Also known as the Buckley Amendment, the Family Educational Rights and Privacy Act (FERPA) of 1974 spoke to the issue of the confidentiality of students’ personal and academic records. Under this Act, personal “directory” information about students such as their name and phone number—even their date and place of birth—may be disclosed without a student’s permission.

FERPA also gave students both the right to inspect letters written about them and the option of waiving access. This waiver of access, in effect, takes away the student’s right to see the letter and protects the recommender’s right to offer subjective commentary, as long as the letters are used “only for the purposes for which they were intended” (9). If over-analyzed, this document seems to create something of an “access tug of war” between student and writer, and it’s unclear which party has the most responsibility to establish access rights, but in practice the standard is clear: Students seeking a letter of recommendation are typically asked to sign a statement waiving their access rights to the letter, usually right on the letter request form itself.

Increasingly, in the context of recommendation letters, some schools interpret FERPA and its amendments to mean that a letter writer cannot reveal certain academic records without written permission, specifically the following:

• the names of courses taken;
• the grade received in a particular course;
• class rank;
• GPA.

For letter writers, of course, such restrictions can be inconvenient or even seem downright silly. As opposed to revealing, say, a student’s Social Security number, which has obvious ethical and financial complications, revealing a student’s grade in a course is one of the best and easiest ways to praise the student. Therefore, many writers simply include such information without concern for FERPA regulations, trusting that there will be no repercussions from the student.

Whatever your past habit has been, the best practice is obviously to ensure that the student has signed a written statement waiving access rights, and schools are increasingly detailing exactly what that statement should include. As one example, the University of Utah provides students with a “Permission To Release Education Record Information” form, including the student name and ID number and the following statements:

I give permission for ______ to write a letter of recommendation on my behalf, and for the purpose of ________. This letter can include the following information:

Please check all that apply:

    □ Grades    □ GPA   □ Class rank

Please send letters of recommendation to: __________.

I waive my right to review a copy of this letter of recommendation now and in the future.

    □ Yes        □ No

As another example, a professor from the University of Minnesota has crafted a "Student Reference Request Consent Form"—a checklist covering everything from the purpose of the reference to permission to provide an evaluation of academic performance to a waiver of access rights.  I like this form in particular because it teaches the student about protocol.

Even if overly detailed for the tastes of some, these approaches do seem prudent in relation to FERPA and how schools are interpreting it. Also prudent is to check with your academic institution as to its specific policy regarding recommendation letter access. This policy might be more obviously aimed at students rather than at faculty, so a bit of digging may be necessary. However, with the policy usually directly tied to FERPA, doing a search for references to FERPA on your school’s website is a good beginning.

In 2003, the Department of Justice investigated a claim by a student at Texas Tech University that religious discrimination was inherent in a professor’s policy of not writing recommendation letters for students who didn’t support the theory of evolution. On the professor’s website, he had told students seeking letters: “I will ask you, ‘How do you account for the scientific origin of the human species?’ If you will not give a scientific answer to this question, then you should not seek my recommendation” (13). The student hadn’t sought a letter from the professor, nor enrolled in his class, but had sat in on the class for a few days and objected to the professor’s posted policy. While a representative at the Liberty Legal Institute, which supported the student's claim, called the professor’s actions “egregious conduct,” a spokeswoman at Texas Tech defended the stance: “Professors don’t have to write recommendations at all, and we certainly don’t tell them who they have to write for . . . He’s not saying he wouldn’t write a letter for a Christian—he’s saying he wouldn’t write a letter for someone who doesn’t believe in evolution” (14). The investigation was dropped after a short time, after some changes to the professor’s website, including a comment that the policy should not be “misconstrued as discriminatory against anyone’s personal beliefs” (13).

Whatever stroke you swim in this ethical soup, you are well-advised as you write letters to consider the issue of discrimination as a complex, potentially combustible one. Any number of stances might be supported in the above scenario, but the back story to this tale should be equally interesting: presumably, the professor came to this stance after repeatedly facing the issue with students requesting letters—i.e., the position developed from experience. When a claim of discrimination surfaces, considerations of experience and intention seem to be critical on behalf of both parties. And these considerations are best weighed within context. In the context of letters of recommendation, writers must be concerned about discrimination based on gender, race, and a host of personal circumstances.

Considering gender, one study of over 1,000 letters of recommendation from the 1970s provides some revealing language. One candidate was described as a “tallish blue-eyed blond,” another was cited as “not neglecting her family,” and another was characterized as having a “remarkable [devotion to scholarship] in a young woman who is physically so slight and so pretty” (15). Such commentary, which baldly smacks of sexism and stereotypes, helps us see the obvious danger of discriminatory language and examples in letters of recommendation.

A successful lawsuit cited earlier involved reference letters and claims of race discrimination (7), and numerous studies have explored the issue of gender-biased practices in letters. An assessment of 300 letters written for medical faculty reinforced gender schema of women as teachers and students and men as researchers and professionals (16), and even studies that have found no substantive evidence of sexism in letters have supported findings of content differences based on the writer’s gender and evidence of gender solidarity (17, 18). To generalize, then, there is an argument that not only do males and females write differently in letters of reference, they write about males and females differently. The bottom line emerging from such academic findings—not to mention from the employment of common sense—is that writers clearly must not make statements in letters that could serve as a basis for discrimination. And a good number of statutes define the various headings under which we must avoid discrimination: race, gender, nationality, sexual orientation, politics, religion, age, appearance, marital or parental status, or any handicapping condition.

But for the letter writer the issue is often highly nuanced: Suppose an African-American professor wants to comment on his student’s role as an officer in the campus Black Caucus? Suppose a professor of women’s studies wants to celebrate her female student’s paper on depictions of feminine stereotypes in 19th century paintings? Suppose a political science professor wants to cite her student bringing experience from his native third world country to classroom discussion? Should these writers avoid such commentary based on concerns of discrimination? The considered answer here is likely “no,” but neither can such comments be made in a way that race, gender, or certain personal characteristics are likely to become an inappropriate criterion in the decision-making process. As noted earlier on the subject of discrimination claims, the writer’s experiences and intentions are highly relevant, as is the letter’s context.

Avoiding Discriminatory Practices in Letters

To avoid discriminatory practices while still addressing appropriate personal characteristics of the candidate, consult the following list of questions:

Is there any good reason to reveal gender, race, or other potentially discriminatory characteristics within the context of the application as a whole? If not, strictly avoid doing so (other than by use of the appropriate gender pronoun, of course).

If race, gender, or other personal characteristics are relevant to the application context, is the student invited to comment in these areas? If so, are you as the letter writer specifically invited to do the same? Noteworthy examples include the Soros Fellowship for New Americans or a scholarship specifically for women in science. Here, effective commentary on nationality or gender within the context of a field might be considered relevant, though such commentary would still have to be concerned with tone and proportion.

Do you have a meaningful affiliation with the student which goes beyond the student/teacher relationship? Do you have personal information about the student that you think is highly useful to mention? Is it naturally relevant? Is a selector likely to find such information automatically helpful and benign or needlessly distracting? Does the personal information lift up and humanize the student or does it reinforce stereotype? Answering "yes" to any of these questions increases the likelihood that the student's race, gender, or other personal characteristics are valuable contributions to the letter's content.

Even with this list of questions, getting near such matters feels like too much of a hot button issue for many writers. When in doubt, some writers actually ask students their opinion of how such details might be addressed if at all, while others consult with colleagues or do some digging to find out their school’s policy. Certainly, all schools have considered this issue, and some such as the University of Alabama in Huntsville—in its “Legal Implications of Letters of Recommendation” (19)—publish guidelines. Amidst this school’s guidelines, for instance, is this example of what a blundering discriminatory statement might sound like: “For a 55-year-old non-traditional student, she has a remarkable record, particularly in view of her inner city background.”

Perhaps the simplest rule of thumb is this: When writing letters, avoid comments that would make a person of sensibilities become distracted enough to wince.

In 2002, the American Academy of Arts and Sciences published a paper arguing that the fundamental means we use to evaluate rising students—by quantitative measures such as grades and qualitative measures such as recommendation letters—have substantially changed over time. Studies on grade inflation covering the years 1960 to 1997 found measurable GPA increases in every institutional type, greater percentages of grades awarded in the A range, and lower percentages of grades falling below the B range (20). The Vietnam War is often cited as a beginning of the grade inflation trend, with faculty awarding students unearned grades in order to keep them from dropping out of school and becoming subject to military service. Other cited factors contributing to this trend include curriculum changes that reflect the changing tenor of the culture and the workplace, the increasing use of student evaluations (with critics arguing that faculty give easy grades to gain popularity), the increasing use of adjuncts, and the greater number of students in college aggressively seeking the grail of post-graduate education. Finally, a trend towards students viewing themselves as “consumers,” with schools as their service providers, also contributes to the problem. One study in 1999 found that large proportions of students in five different courses saw grade inflation as the norm—i.e., even students who self-reported doing ‘average’ work still expected Bs or As (21). Put simply, many students expect to be given high grades, even when delivering mediocre performance.

Certainly, the issue of grade inflation has drawn significant academic attention, and in some cases even new policy initiatives, over the last few decades. Since 2000, the schools that have introduced policies to curb grade inflation include Columbia University, Harvard, the Washington University School of Law in St. Louis, and Princeton University. Princeton University’s policy, in particular, drew much attention in 2004, with the Dean of the College Nancy Malkiel citing grade inflation as “an intractable national problem” and calling for faculty “to give students the carefully calibrated assessment they deserve of the quality of their course work and independent work” (22). The Princeton proposal called for programs and departments to ration grades more carefully, awarding less than 35% of grades in the A range, and less than 55% in the A range for independent work by juniors and seniors. The proposal gained almost 2/3 support in a faculty vote (23). Five years after the policy was implemented, in 2009, Princeton reported marked progress in curbing grade inflation, with about an 8 percent drop in the number of As awarded. A statement issued by the Faculty Committee on Grading noted that “These results confirm once again that with clear intent and concerted effort, a university faculty can bring down the inflated grades that—left uncontrolled—devalue the educational achievements of American college students” (24).

Meanwhile, a trend parallel to grade inflation is evident in letters of recommendation written by faculty, whether written for students or for peers. Both students and faculty expect and sometimes pressure their recommenders to write glowing evaluations, candor is replaced by gloss, and qualitative distinctions become blurred. As one writer puts it, commenting on letters written for tenure and promotion candidates, “Puffery is rampant. Evasion abounds” (25).

As discussed in Sissela Bok’s fascinating book, Lying: Moral Choice in Public and Private Life, the act of writing an inflated recommendation has become sanctioned simply by pervasive practice. The writer’s usual reasoning is this: “It helps someone, while injuring no one in particular, and balances out similar gestures on the part of many others” (26). In fact, most evaluators feel that if they do not conform to an “inflated set of standards” within a “system where all recommendations are customarily exaggerated” (26), they do their students unintentional harm.

Though the practice of writing inflated letters of reference is harder to quantify and less studied than the issue of grade inflation, the backlash trend against inflated letters of praise seems to be taking the same path: Scholarship jurors are urging more candor that will provide clearer distinctions among candidates, selection committees are calling for more context about program standards so they may better assess a candidate’s worthiness, and letter reviewers are growing more vocal about their need for specificity and credible information. In short, readers want informative letters they can trust.

What then are the best practices we should use amidst these trends, while acknowledging the student’s need for an effective, helpful letter of recommendation? While recognizing that we work in an academic culture where inflated letters are common, our best practices include considerations such as praising (even criticizing) in ways that lend credibility, and understanding the audience’s need for a letter that informs in a way that provides a best fit between candidate and opportunity. The challenge and charge is to reduce inflationary rhetoric while still honoring and serving the student’s needs.

On the topic of praise in recommendation letters, paradoxes commonly arise. The same person recommends three different candidates for the same scholarship, and each one is redundantly put forth as “absolutely the best student I have ever had.” Every student within a pack of 50 is lauded with the superlatives “outstanding” and “excellent,” yet their GPAs and GRE scores range from perfect to average. And how about the statistical conundrum when 60% of the reference letters from a particular school claim that the recommended students from this year are in the top 5% of students the school has ever graduated? Is this year’s crop really that good? Reacting to this trend of over-the-top praise in letters even in the mid-1960s, one researcher facetiously titled his journal article on the subject, “Mine Eyes Have Seen a Host of Angels” (27).

Unsurprisingly, seasoned selectors can become tone deaf to hyperbole, especially when it’s presented ineffectively—they’ve heard it all before, and unqualified glorification with no supporting evidence or contradictory evidence is simply not credible. We have a system where praise is given in such lavish heaps in recommendation letters that it ironically becomes suspect. As one study put it, search committee members become frustrated with “the contradiction presented by a volume of glorious recommendation letters and a candidate’s weak academic record and a disappointing personal interview” (28). One professor at the University of Illinois at Chicago asserts that, in economics, “it is well known that a certain Nobel laureate claims in every recommendation he has written that the present recommendee is ‘the best student I have ever had’ . . . since around 1953” (29).

One impact of this trend, in cases where selectors actually take the time to follow up and check on the veracity of a claim, is that the evaluation process becomes driven through “increasingly informal channels” (20), where people simply call or e-mail the recommender for the “real information,” resulting in what could be either a more thorough or a more relaxed process, depending on the ethic of all parties involved.

There is also a forceful argument that hyperbolic praise in recommendation letters does harm both to academia and to individuals. One philosopher posits that the practice “obviously injures those who do not benefit from this kind of assistance; and it injures them in a haphazard and inequitable way” (26), with two equally qualified candidates rated differently based on their recommenders’ different levels of flattery. A 2002 report assembled by academicians from different fields finds that an overly laudatory letter “cheats those excellent candidates who deserve great praise and gives less distinguished applicants an unfair and unearned advantage. It may also cause the employer or educational institution to have unrealistic expectations of the candidate” (20). As one instructor at Harvard succinctly puts it: “By rewarding mediocrity we discourage excellence” (30).

The argument that inflated praise is desirable or necessary in letters can also be a powerful, student-centered one. Respondents defending the practice in one study argued that students who are merely satisfactory in college sometimes go on to do very well, so they might be given the benefit of the doubt, and that many students change and improve with experience (8). Knowing that exaggeration is the norm in letters, many writers practice it simply to avoid hurting their students. As one professor of law and philosophy at the University of Texas at Austin noted: “Someone who is candid risks damaging their students, because candor is uncommon” (25). When we write a reference letter, some say, we are responding to the best in the student, and projecting what that student may become in time.

Clearly, what is needed is a form of praise that honors both the student and the letter evaluator equally. A 2002 study by the American Academy of Arts and Sciences suggests a dual ethical responsibility to the candidate and to the evaluators making the decisions. On serving the needs of the evaluators: “A rephrased Golden Rule is the best guide: Write to others the kind of letter of recommendation you would like to receive from them. To follow the rule is responsible professional conduct. Not to follow the rule perpetuates harmful practices in the academy” (20). Though by definition a recommendation letter will always be complimentary and flattering, recommenders serve their students and academia best by writing a letter where praise is measured and exacting, where superlatives are backed up by demonstrative examples, and where statistics about student ranking or quality are used with consistency and great care.

Given the cautions provided in the previous sections against inflationary rhetoric and hyperbolic praise in letters, some attention to the role of criticism—or perceived criticism—is critical.

One study from the field of psychology—where one must assume that honest assessment is held in high value—is instructive. In a large survey where 98 percent of the respondents were doctoral-level psychologists evaluating students for potential clinical practice, a “nontrivial minority” of respondents said they would not include negative evaluation of students even with knowledge of negative behavior. Specifically, 12.4 percent said they would exclude mention of problems of alcohol or drug abuse, and 43.2 percent said they would not mention anxiety or depression. In the same study, a sample of 116 psychologists who had recently read reference letters said that negative characteristics were infrequently described. Readers of letters in this study assessed that “writers feel more obligated to the student than to the letter recipient” (31).

Certainly, the face-to-face factor plays a strong hand in this dynamic. It’s difficult to look at a student, agree to write a letter, then feel we’ve turned our back on the student by articulating criticism. We also know that letter evaluators actually read with a great deal of sophistication and subtlety when trying to sniff out negative comments (in contrast, praise is read rather simplistically, one could argue). As an example, one study found that negative comments in letters could be grouped into five categories: relative progress (“he’s come a long way”), disadvantaged background (“she’s overcome cultural obstacles”), explicitly negative (“his work is competent but not distinguished”), remediative (“improvement is needed”), and inconspicuously ambiguous (“she is aware of her strengths and weaknesses but won’t take on things she’s unqualified to do”) (32). With such possibilities open to broad reader interpretation, we have to assume that even a minor comment can be interpreted and remembered as definitive and potentially damaging, even in an otherwise positive letter.

Red Flags and Omissions

Plenty of writers—intentionally or otherwise—wave red flags to their readers as they write letters. These red flags might take the form of distancing (with the writer taking pains to show limited knowledge of the student), a critical incident ambiguously offered (a student being called initially aggressive, then “winning over” the writer) (32), or the presentation of so many irrelevancies (either about the student or the recommender) that readers feel there is something to hide. One study published in the journal Academic Emergency Medicine tied the issue of veracity directly to authorship: “Veracity requires that authors avoid sins of omission and commission, understatements as well as overstatements” (12).

One of the loudest ways some writers criticize, or seem to, is through omission. Studies in the medical field in particular show that there is a finely tuned radar for what is not said in a letter. If one has a sense that the writer kept mum about important information, one assumes that something negative is lurking between the lines. “One of the most challenging features of letters of recommendation for medical faculty is the growing tendency not to state the negative, but merely to fail to state the positive” (17). In fact, in one study of letters for surgical residencies, “The commonly used phrase, ‘If I can provide any additional information, please call . . .’ was almost uniformly identified as a strong negative comment” (33). Such red flags and omissions don’t always exclude a candidate, of course, but they can do so in the hands of a hasty reviewer, or they can require readers to take extra time with an application to sort through the writer’s intentions.

Cross-Cultural Factors

An associate professor at Duke University once “pumped up the volume” in a letter sent to a university in Great Britain, calling a student “outstanding.” Soon he received a call from the search committee, asking if the letter had been forged. “It was so hyperbolic in their eyes that they couldn’t believe it,” the professor said (25). He found what many have described anecdotally—that British evaluators lend more credibility to a letter that is not inflated, and even includes at least one criticism. This is especially relevant when one knows that a British evaluator will be part of the process, as with the Marshall and Rhodes scholarships.

One study that compared recommendation letters from four countries found that the letters all seemed to have the same purpose, but that “the ways that writers from different cultures express support vary a great deal” (34). Another study involved only a small statistical sample, but the researchers noted that “even letters from Canada were less hyperbolic than those from the USA” (16).

Reasons and Ways to Criticize

Despite all the concern raised here about the damning power of criticism, we must remember to keep perspective and recognize that the mentor/student relationship often does involve honest critique. I’m reminded of my early days as a fiction writer, where I gained the most help from one challenging mentor (who was, perhaps significantly, British) who wrote at the end of a clever but vacant piece of mine, “You’re very clever, now why don’t you write me some fiction?” I am also reminded of numerous examples where I have given honest criticism about students verbally during interviews or background checks, and in writing amidst an otherwise positive letter, and those students still landed the desired opportunities. My hope and belief is that the criticism lent more credibility to my overall evaluation, and that the evaluators put it into the proper context.

To criticize artfully and kindly when writing letters, consider these practices:

1. See if the application materials call for criticism. Many times, a statement will be provided to the referee also inviting your assessment of a candidate’s weaknesses. Take this as a sign that careful critique is desirable, and even cite that statement as you give the critique. Consider this especially when there are cross-cultural factors at work, or when you’re recommending a student in a field with a strong sense of hierarchy.
2. Limit your criticism to one paragraph (probably late in the letter) rather than pepper it throughout, and be direct and affirmative as you offer it rather than ambiguous, avoiding comments that seem to be veiled criticisms. Even phrases such as “To the best of my limited knowledge” or “I suspect that” could be read as negative no matter what praise follows. Better to say “his research skills are not yet proven with lab experience” or “her teaching could be improved with a higher level of confidence.”
3. Avoid highly negative comments (looked on with as much suspicion as exaggerated praise), hedges, unexplained asides, and irrelevancies.
4. Discuss your criticisms with the student, note the response you get, and indicate right in the letter that this discussion and response took place.
5. Openly tie your criticism to your sense of ethics as a letter writer (as detailed in the next section). Define yourself as a holistic evaluator.

Among the many mini-lessons in ethics offered up by The Dick Van Dyke Show (including not eavesdropping on your neighbors, not sticking your big toe in the bathtub faucet, and perambulatory care around ottomans), episode #108 is a moral tale about writing recommendation letters. Rob Petrie’s old pal, Neil Schenk, asks Rob to write him a positive recommendation letter for a job, playing on Rob’s guilt about owing him an ancient favor. The hapless Rob waffles, feeling he can only write him a neutral letter at best, and when Rob hesitates Neil simply writes his own letter and manipulates Rob into signing it. In the end, of course, the morally straight Rob Petrie opts for telling Neil Schenk and the employer who received the inaccurate letter the truth, and because it’s the world of the sitcom their old friendship is nevertheless preserved.

The dilemma above is not uncommon in the academic world, though here the situation is reversed, with the faculty member sometimes inviting the student to initially draft his or her own letter for review. Though some faculty swear they’ve never heard of this and some see it as a wholly unethical practice, I’ve talked to a good number of faculty who say a department head once asked them to write their own reference letter draft, and I’ve worked with plenty of students who were asked to do the same by a faculty member, who usually cites a busy schedule as prohibitive to the task. Presumably what follows is a document that is reviewed and rewritten, and hopefully the ghostwritten draft isn’t so much a letter in sentence and paragraph form as a list of accomplishments or examples that are useful to include in the letter, which is ultimately written by the proper party.

Without overstating the case that the act of writing recommendation letters has ethical implications and repercussions, there are a number of areas where questions of best behavior and protocol come to mind. Especially if a pushy student tries to control the process, it is worth remembering that the key decisions are yours to make, including whether or not you write the letter, your dual obligations to the student and the letter evaluators, and whether or not you share the letter with the student.

Saying No

Sometimes the kindest, most responsible thing we can do for a student is refuse to write a letter of reference. Most faculty try to warm the student to this idea subtly, suggesting that they are too busy, that they’re not sure what to say, that there must be others who can write a more positive letter, or that they simply don’t know enough about the student, whom they might have had in class years ago.

Savvy students will usually recognize such responses as the faculty code for “no,” but the savviest might offer you a resume and a meeting or an e-mail to help you generate positive detail for a letter (a few have won me over by doing so). Also, the pushiest and most histrionic students might insist that you really are the best recommender they’ve got, and that your letter is critical to their very lives.

In this case, assuming you still don’t want to write a letter, here are some good reasons for not doing so, which you might share with the student if so moved:

• Your limited support of the student doesn’t match the weight of the opportunity—as in the case of a national scholarship—and your neutral or unsupportive letter might only do the student unintentional harm.
• You really know nothing beyond the student’s recorded grades, and you’re not inclined to get to know more, for reasons of time or temperament.
• You genuinely dislike the student, for good reason, perhaps because of bad classroom behavior such as chronic lateness or missed deadlines, or because you know the student abuses drugs or commits illegal acts.
• You’ve had encounters that indicate deep depression, unusual personal problems, or a level of anxiety that would affect the student professionally.
• You have knowledge that the student cheated on an exam or plagiarized on a paper, and thus you can’t justify helping to advance the student’s career.
• The student approaches you and nags you in such an unprofessional manner that you are fully persuaded that your endorsement would be a lie.

Admittedly, some of these reasons do sit you in judgment, and many faculty wouldn’t feel comfortable sharing their true reasoning with students, but good students always have options about whom they can ask for letters, and poor students and students in trouble sometimes do need to face the reality of how they got there. Faculty who say no for good reason suffer less than those who agree begrudgingly to write a letter, then feel divided in the very act of giving praise.

Gatekeeping

In the 1970s, the Director of the Hastings Center of the Institute of Society, Ethics and the Life Sciences, having problems with a new hire, contacted the employee’s references and took them to task. The references sheepishly admitted to the worker’s bad history, but felt obliged to help him. “Surely we owe something to our profession or field,” the Director argued—“that its standards be high, that it not tolerate the inept, much less outright malpractitioners, and that it work to serve the general good rather than the self-interest of its practitioners” (35).

In a few extreme cases, former employers were sued, having provided positive references for an employee despite concrete knowledge of past illicit or illegal behavior, and when the same behavior surfaced in the new workplace, the former employers were held accountable for not being forthcoming in their letters (8, 19).

In the rare cases in academia when someone follows up with the references after awarding or hiring someone who turns out to be a dud, the same story is often told—we were aware of the problem, but didn’t want to harm the person’s chances by bringing it up. What surfaces is something of a “buyer beware” mentality, as one writer calls it (27), with the onus on the recipient to deal with the problem. But many feel strongly that it should sometimes be the letter writer’s job to serve as a protective gatekeeper. As one professor at Florida State University put it: “I write letters . . . to advance the academy” (1). A respondent in one study summed up the sentiment about the need for gatekeeping in this way: “I know that our university program is represented by our graduates working in the field. So I do not wish to lesson our ethos, and neither do I want to cause problems for either employer or graduate by putting people in positions for which I have no evidence they would do well” (8). In medical fields, where letters of recommendation are critical tools for deciding between candidates, the letter writer might feel a stronger sense of responsibility to the medical profession than to the individual candidate. As stated in a journal article from Academic Emergency Medicine, for the letter writer “there is an implied duty to future students, colleagues, researchers, and patients who might come in contact with the applicant” (12).

In practice, much of this issue simply goes to writing a letter where praise and criticism are used fairly, or just saying no to writing a letter in the first place—issues discussed earlier—but there may be cases where you write from the position of protecting your school’s or profession’s reputation. I once wrote a negative letter for a student applying to grad school in my own program, but only after the student repeatedly refused to take no for an answer, after warning the student that my letter would not be wholly positive, and with certainty that this student was a terrible fit for the program. My loyalty to our program became the most important factor in deciding to write a negative letter. Often, gatekeeping is far less extreme, simply taking the form of passing on specific, credible concerns, but it usually involves some sort of mental wrestling match as you decide where your loyalties lie most. The more allegiance one feels to those receiving the evaluation, the more the writer might feel required to serve as gatekeeper.

Perhaps the most ethical principle is this: If you criticize with the intention of upholding the standards of the profession, you should make that motivation clear in the letter, and still refrain from criticizing irrelevantly or out of proportion.

Being Honest with the Student

As I write letters for my students, I believe in partnering with them on the process, beginning by asking them their opinions of their accomplishments and shortcomings. Usually, they don’t praise themselves enough (though sometimes they overdo it, and I must provide a more mature perspective), and they are highly capable of articulating their own weaknesses. Such a discussion fosters honesty between us, and I find I can thus write a more even-handed letter, sometimes including the student’s self-reflection and self-criticism as part of my text. Also, if I feel that I must write a limited or neutral letter for the student, I can admit that more easily if we’ve already had an open discussion about the process.

When writing letters of reference, one of the best ways you can smooth the process for yourself and foster honesty with the student is by developing a set of practices and policies, being up front about how you feel about writing letters and your personal protocol. Here are excerpts from two professor’s policies:

“If you have made no impression in my class, it is unlikely that you will get a good letter” (36).

“You need to have completed at least two of my courses. . . . You need to have earned a grade of B or better in all courses taken from me, received an A in one of my courses, and have at least a 3.0 GPA overall” (37).

“Ask me if I have the time to write the letter and if I would feel comfortable writing a supportive and positive recommendation letter for you. . . . I would rather decline writing you a recommendation letter than to write you a vague or irrelevant one” (38).

Note how these policies anticipate common problems in the process, set forth the professor’s ethic for how quality is defined, and build in the possibility of a negative letter or no letter at all. By developing such policies and being up front about them, these professors help students realize the weight given to recommendations, and underscore the boundaries of the professor’s protocol—which certainly might be stretched for the right student, one would assume, but not without the student recognizing the rules of the game.

Being Honest with the Letter Evaluator

Beyond the need to be honest about your assessment of a student’s abilities, there are other areas where authors should uphold an ethic of honesty, in particular in matters of authorship, criticism in your letter, and letter confidentiality.

In some cases, letters may have multiple authors or be co-signed, and if such matters are ambiguously presented or only gleaned through the letter’s signature line, letter evaluators are likely to be puzzled, and the letter’s credibility could come into question. At Harvard, teaching fellows who author a letter are invited to prepare a draft and share it with a course instructor, who then rewrites or simply co-signs it (38). In such cases, a carefully worded statement defining the chain of authorship is appropriate in the text of the letter, perhaps even right up front. I’ve also seen cases where a professor endorsed a letter written by a grad student by providing an imbedded or separate signed statement, noting that the letter had been read by the professor and met with his or her approval.

Some writers, wanting to make sure that evaluators recognize why criticism is being offered even amidst full endorsement of a student, include statements about their own ethic when writing recommendation letters. One study suggests “A written qualifying statement acknowledging the tendency of recommendation letters to overstate candidates’ desirable qualities, and stating the writer’s intention to avoid this trap by providing a more complete letter covering both strengths and weaknesses” (39). The intention here, of course, is to lend credibility and context to your honest and complete assessment of the student.

Finally, some writers have strong feelings about whether or not a student should be privy to a letter’s contents, and some studies recommend that you “state in the letter whether or not it was shared with the applicant” (31). The effect of such a statement can go in either direction, in that some feel a confidential letter shouldn’t be shared with the student under any circumstances, thinking that it destroys the impact of confidentiality; others feel that the student should be highly involved in the process of generating detail for the letter, and thus the issue of confidentiality for them is much more flexible. If you do make such a statement, again your concern should be with making it without equivocation. To demonstrate, note how there is a great difference in how these three statements might be interpreted:

“Per the usual protocol, I have not shared this letter nor its contents.”

“After a lot of thought, I decided that John might be really upset if he saw this letter, so I’m keeping it confidential.”

“After she signed the right of access waiver, I shared portions of this letter verbally with Janet, respectful of her maturity in handling my evaluation.”

If no such statement is made, a confidential, non-shared letter should be assumed.

Sharing the Letter with the Student

Through both serendipity and design, I’ve seen most of the recommendation letters written for me. In addition to the scenario I discuss in the Preface of this manual, several of my professors and my dean have given me what are commonly called “blind copies” of my letters either before or after sending them and after I had waived my access rights. I’ve talked to many others who have also seen at least one recommendation letter they’ve had written for them, either because they were involved in the process of helping to generate detail for it or because the writer had a policy of sharing even confidential letters.

In reading my own letters and considering the process carefully, I’ve developed a practice of sometimes sharing a letter or portions of it with students, but I typically do so after the decision based on the letter has been made, and when the decision was favorable to the student. Though never obliged to share letters with students who have waived access, many faculty develop some sort of access policy. If you choose to do so, the safest practice is to consider each case individually, based on the detail of your letter, your level of criticism, the status of the letter at the time of access, and the student’s maturity.