E
Emil Chuck, PhD
Advertisement - Members don't see this ad
Check out this article or other pre-physical therapy articles at Student Doctor Network.
Do you remember the first time a student asked you to write a strong letter of recommendation? I remember the student who asked 10 days after I had already committed to writing 15 other letters. The student bargained further in the hope that I would make an exception, but the other 15 requests had similar deadlines. Medical school application season was nigh, and everyone wanted their letters submitted before June 1, right after the registrar’s deadline to submit my final semester grades.
I hardly knew these students, aside from their midterm and final grades in my 200-person lecture class. Everyone earned A’s or B’s, so they felt they were doing well enough to avoid asking for my help during office hours. Each one sent me a customized email with an upload link.
I planned to begin writing during finals week, but my follow-up emails to these students went unanswered (which I attributed to their focus on their exams). But when finals week was over, I emailed each student to ask for more information so I could better highlight their strengths and fit for their desired graduate school opportunities.
All the medical school applicants sent me a link to the Guidelines for Writing a Letter of Evaluation for a Medical School Applicant.1 However, I often could not recall information about each candidate; some of us use an online filing system to organize documents, while others hope our search function will find the candidate’s emails.
While we expect our lecturers to effectively manage introductory courses with hundreds of students, we rarely teach new faculty how to write a meaningful letter that helps a medical, veterinary, or dental school applicant, including those who need support letters for summer research opportunities. This guide is designed to help you write effective letters of recommendation help your students achieve their dreams.
Admissions committees cannot condone the practice of having candidates write reference letters that the reference signs off on. For professor letters, the practice violates the principle that a student waives rights to review the letter under the Family Educational Rights and Privacy Act (FERPA), as discussed below. Instead, admissions committees expect references to maintain their professional responsibilities to deliver a supportive letter of recommendation.
This practice also invites candidates to use generative AI tools to ghost-write letters of recommendation. Admissions committees are highly concerned about references written predominantly by generative AI, and many schools may draft guidelines for writers about appropriate AI use.
Experienced admissions committee members develop preferences about the elements of an effective letter of recommendation because they often read letters from the same references over many application cycles. Many letters are effective because the writers organize information about an applicant’s academic performance, personal growth, professional demeanor, and communication skills in ways that would be valuable to the future professional community. The most effective letters reaffirm the applicants’ purpose and goals as future professionals and add to the application essays and interview responses.
Recently, admissions officers have boldly expressed concerns about the growing prevalence of generative-AI recommendation letters. Many commented on this topic from the 2026 conference of the National Association for Advisors of the Health Professions:
Admissions committees expect references to use genAI tools appropriately when outlining content, correcting grammar or spelling errors, auditing/reducing inappropriate bias, or adjusting for professional clarity. Generative AI tools should not be used to create generic reference letters, invent encounters with the applicant, or wordsmith to express a stronger endorsement when the evidence is lacking. Higher education institutions are also concerned about the use of genAI when faculty members are using tools that may not properly protect student information from larger public language models.
References should be informed that health professions programs have adopted different policies on using genAI in letters of recommendation.2 They may be required to attest to the use of genAI (2026-27 CASPA policy).3 Similar policies are being considered in graduate medical education for letters supporting residency applicants.4
Writing letters for students is an uncompensated faculty responsibility. Faculty are not promoted based on the number or quality of reference letters they write for students. While your letter authenticates your official capacity representing your department, school, or university, most faculty letters do not require approval from the chair, dean, or university counsel.
In your syllabus or when making your first announcements to your class, tell your students when you are accepting requests to write letters, the circumstances for requesting your help, and the requirements for a strong letter. I needed at least 2 months to respond to each letter request.
Because many students may ask for letters from those who previously took your class, meetings with these students should be scheduled separately from your designated class office hours. Some evaluators ask students to meet at a nearby cafe or coffee shop for a more casual conversation; I know of another university whose student government underwrites vouchers that let students take a professor to lunch or dinner on the government’s tab (no alcohol or unnecessary indulgences). Otherwise, meetings can be done over a webcam or a leisurely walk around campus.
I find it helpful to get the online description of a scholarship, internship, or research opportunity, so students should provide this information before meeting with the reference. The meeting should go over why the student wants the professor to write the recommendation and specific “talking points” that put the student in a strong position.
As someone who reviews letters of recommendation, I want to know the following:
As a reference, if you know your answers will likely not support the applicant, you should decline by saying you are unable to write a “strong” letter of recommendation.
For me, my best letters come after at least two meetings over six months with the student that discuss their personal journey, their inspiration to pursue a health professional career, and whether they have leveraged the university’s resources (prehealth advising, career services, learning support, multicultural or interfaith centers). An actual packet with a resume, a biosketch, a photo for the application process, and a shortlist of desired programs, especially a printed copy, helps me craft a highly individualized letter of support.
One-on-one meetings also give me insight into how the students will perform in conversational interviews. What stories about their successes and failures can they comfortably share or discuss? How do they talk about their study habits and time management?
Many faculty peers who engage in the admissions process often ask for “code words” that signal some comparative value: for example, what is the difference between “highly recommended” and “strongly recommended”? I have developed a working rubric with appropriate phrases to explain my summary recommendations, highlighting how the candidate has earned that assessment based on what I know about the candidate’s performance and connection with students in my course. Sometimes I share this rubric with the candidates in advance.
However, some evaluation portals ask you to complete their own rubric. The most recent ADEA AADSAS evaluation portal (2026-2027 cycle FAQ5) asks us to provide a Likert score for 10 attributes about an applicant.
Likert Levels:
Even if the applicant is not interested in dental school, I insist on meeting with each student to go over this rubric (or my own) and identify areas where I can offer insight and comment. I know most students would not ask me as a reference if they thought they were not ready in any one of these competencies; more likely, I have been asked because they feel I can articulate how the student exceeds expectations in all ten. But if the student is in my large lecture class, I couldn’t assess them as “exceeding expectations” in integrity. If the student has worked in my lab, I feel more comfortable rating them on their academic competencies. My expertise as a professor gives me greater insight.
The Student Doctor Network Review article How to Assess Your Academic Competencies, Student Doctor Network Review includes an example academic competency rubric following AAMC definitions.6
As a professor, I should be familiar with the Family Educational Rights and Privacy Act of 1974 (FERPA).7 Faculty guidance is likely available at your institution (see FERPA Release for Letters of Recommendation & References at Towson University).8
Because students waive their right to review my recommendation letters, I have always sought transparency in my work with students. Over the years, I have shared an outline of my recommendations. The length of the letter varies with the depth of my mentoring relationship with students, but I keep the length of my letters to within two pages as follows:
Even if I know they have waived their FERPA rights, I share a draft of my recommendation with the student to ensure I haven’t made any mistakes (such as improper naming or gender-appropriate language/pronouns). Those detail errors are noticed by application reviewers and can undermine the candidate’s standing. The student is also assured that their journey to the profession is accurately documented.
The Office of Medical Education at the University of California San Francisco provides guidance for clinical faculty when documenting student performance, including examples of strong recommendations.9
In the past, many students asked former professors to write letters of recommendation, but the schools contacted the students because those letters were not submitted on officially branded letterhead. At my institution, letterhead is restricted to official business for current employees. For those leaving the institution, I advise my peers on our letterhead policy.
Many graduate programs are more accustomed to “digital signatures” that we affix to our letters. A few of us still keep a copy of our signature or have created an artistic “signature logo” to paste into our letters. The rest just type /* their name */ (within slashes and asterisks). At any rate, admissions professionals want to see a signature, and it doesn’t take much effort to satisfy this expectation.
I love learning about these students and their dreams through this process. I am reminded how inspirational these students are in the way they speak about their futures, even as I have to set aside about two to three hours per letter. I hope this guide gives you an opportunity to see how you have guided them towards their dream professions.
I make sure all of these items are fulfilled when I write a reference letter.
Timeline/Deadline
Conferencing Times
In your letter, show your credentials
State how you know the student (drop-down questions in AADSAS)
Who needs the recommendation (i.e., Who is the Audience)?
My observations about the student (refer to the rubric)
Here are suggestions for maintaining your voice as an author and recommender without allowing genAI tools to dilute your evaluation or support.
This article was peer-reviewed by and includes suggestions by Nancy Lorenzon, Teaching Professor, Department of Biological Sciences, Director of Pre-Health Education and Advising, University of Denver.
Heather Brod’s Writing Effective External Evaluation Letters inspired this article.10
No conflict of interest reported.
Generative AI tools were used to support this manuscript in the following ways:
Continue reading...
Do you remember the first time a student asked you to write a strong letter of recommendation? I remember the student who asked 10 days after I had already committed to writing 15 other letters. The student bargained further in the hope that I would make an exception, but the other 15 requests had similar deadlines. Medical school application season was nigh, and everyone wanted their letters submitted before June 1, right after the registrar’s deadline to submit my final semester grades.
I hardly knew these students, aside from their midterm and final grades in my 200-person lecture class. Everyone earned A’s or B’s, so they felt they were doing well enough to avoid asking for my help during office hours. Each one sent me a customized email with an upload link.
I planned to begin writing during finals week, but my follow-up emails to these students went unanswered (which I attributed to their focus on their exams). But when finals week was over, I emailed each student to ask for more information so I could better highlight their strengths and fit for their desired graduate school opportunities.
All the medical school applicants sent me a link to the Guidelines for Writing a Letter of Evaluation for a Medical School Applicant.1 However, I often could not recall information about each candidate; some of us use an online filing system to organize documents, while others hope our search function will find the candidate’s emails.
While we expect our lecturers to effectively manage introductory courses with hundreds of students, we rarely teach new faculty how to write a meaningful letter that helps a medical, veterinary, or dental school applicant, including those who need support letters for summer research opportunities. This guide is designed to help you write effective letters of recommendation help your students achieve their dreams.
Do not ask the candidate to write their own reference letter.
Admissions committees cannot condone the practice of having candidates write reference letters that the reference signs off on. For professor letters, the practice violates the principle that a student waives rights to review the letter under the Family Educational Rights and Privacy Act (FERPA), as discussed below. Instead, admissions committees expect references to maintain their professional responsibilities to deliver a supportive letter of recommendation.
This practice also invites candidates to use generative AI tools to ghost-write letters of recommendation. Admissions committees are highly concerned about references written predominantly by generative AI, and many schools may draft guidelines for writers about appropriate AI use.
Using templates and GenAI can produce less effective letters of recommendation, and admissions committees can tell.
Experienced admissions committee members develop preferences about the elements of an effective letter of recommendation because they often read letters from the same references over many application cycles. Many letters are effective because the writers organize information about an applicant’s academic performance, personal growth, professional demeanor, and communication skills in ways that would be valuable to the future professional community. The most effective letters reaffirm the applicants’ purpose and goals as future professionals and add to the application essays and interview responses.
Recently, admissions officers have boldly expressed concerns about the growing prevalence of generative-AI recommendation letters. Many commented on this topic from the 2026 conference of the National Association for Advisors of the Health Professions:
Admissions committees expect references to use genAI tools appropriately when outlining content, correcting grammar or spelling errors, auditing/reducing inappropriate bias, or adjusting for professional clarity. Generative AI tools should not be used to create generic reference letters, invent encounters with the applicant, or wordsmith to express a stronger endorsement when the evidence is lacking. Higher education institutions are also concerned about the use of genAI when faculty members are using tools that may not properly protect student information from larger public language models.
References should be informed that health professions programs have adopted different policies on using genAI in letters of recommendation.2 They may be required to attest to the use of genAI (2026-27 CASPA policy).3 Similar policies are being considered in graduate medical education for letters supporting residency applicants.4
Set conditions for letters of recommendation for current students.
Writing letters for students is an uncompensated faculty responsibility. Faculty are not promoted based on the number or quality of reference letters they write for students. While your letter authenticates your official capacity representing your department, school, or university, most faculty letters do not require approval from the chair, dean, or university counsel.
In your syllabus or when making your first announcements to your class, tell your students when you are accepting requests to write letters, the circumstances for requesting your help, and the requirements for a strong letter. I needed at least 2 months to respond to each letter request.
Because many students may ask for letters from those who previously took your class, meetings with these students should be scheduled separately from your designated class office hours. Some evaluators ask students to meet at a nearby cafe or coffee shop for a more casual conversation; I know of another university whose student government underwrites vouchers that let students take a professor to lunch or dinner on the government’s tab (no alcohol or unnecessary indulgences). Otherwise, meetings can be done over a webcam or a leisurely walk around campus.
Ask for talking points and documentation
I find it helpful to get the online description of a scholarship, internship, or research opportunity, so students should provide this information before meeting with the reference. The meeting should go over why the student wants the professor to write the recommendation and specific “talking points” that put the student in a strong position.
As someone who reviews letters of recommendation, I want to know the following:
- How do you know the applicant (in what role)?
- How long have you known the applicant?
- How well do you know the applicant (to observe and evaluate their successful attributes)?
As a reference, if you know your answers will likely not support the applicant, you should decline by saying you are unable to write a “strong” letter of recommendation.
For me, my best letters come after at least two meetings over six months with the student that discuss their personal journey, their inspiration to pursue a health professional career, and whether they have leveraged the university’s resources (prehealth advising, career services, learning support, multicultural or interfaith centers). An actual packet with a resume, a biosketch, a photo for the application process, and a shortlist of desired programs, especially a printed copy, helps me craft a highly individualized letter of support.
One-on-one meetings also give me insight into how the students will perform in conversational interviews. What stories about their successes and failures can they comfortably share or discuss? How do they talk about their study habits and time management?
Code phrases and rubrics
Many faculty peers who engage in the admissions process often ask for “code words” that signal some comparative value: for example, what is the difference between “highly recommended” and “strongly recommended”? I have developed a working rubric with appropriate phrases to explain my summary recommendations, highlighting how the candidate has earned that assessment based on what I know about the candidate’s performance and connection with students in my course. Sometimes I share this rubric with the candidates in advance.
However, some evaluation portals ask you to complete their own rubric. The most recent ADEA AADSAS evaluation portal (2026-2027 cycle FAQ5) asks us to provide a Likert score for 10 attributes about an applicant.
- Critical thinking
- Didactic knowledge
- Integrity
- Interpersonal skills
- Maturity
- Motivation (to pursue dentistry)
- Organizational skills
- Professional demeanor
- Reaction to criticism
- Self-awareness
Likert Levels:
- Not observed
- Not ready at this time
- Meets expectation
- Exceeds expectation
Even if the applicant is not interested in dental school, I insist on meeting with each student to go over this rubric (or my own) and identify areas where I can offer insight and comment. I know most students would not ask me as a reference if they thought they were not ready in any one of these competencies; more likely, I have been asked because they feel I can articulate how the student exceeds expectations in all ten. But if the student is in my large lecture class, I couldn’t assess them as “exceeding expectations” in integrity. If the student has worked in my lab, I feel more comfortable rating them on their academic competencies. My expertise as a professor gives me greater insight.
Academic Competencies Rubric
The Student Doctor Network Review article How to Assess Your Academic Competencies, Student Doctor Network Review includes an example academic competency rubric following AAMC definitions.6
Understanding FERPA (for professors)
As a professor, I should be familiar with the Family Educational Rights and Privacy Act of 1974 (FERPA).7 Faculty guidance is likely available at your institution (see FERPA Release for Letters of Recommendation & References at Towson University).8
Transparency and Trust
Because students waive their right to review my recommendation letters, I have always sought transparency in my work with students. Over the years, I have shared an outline of my recommendations. The length of the letter varies with the depth of my mentoring relationship with students, but I keep the length of my letters to within two pages as follows:
- The opening focuses on the purpose of the reference letter and introduces the candidate. I strongly recommend [candidate’s name, application ID if available] as [professor/supervisor]… as a participant in your summer enrichment program.
- I may briefly add more context about the size of the lecture class or any additional opportunities to mentor the candidate.
- Summary recommendation with two or three specific accomplishments that show the candidate’s growth, maturity, and trustworthiness. I will specify skills, techniques, or projects that support my observations.
- Envision the candidate’s success with a future opportunity. I present a closing statement commending the candidate as an appropriate fit for a scholarship, internship, or job opportunity, based on the information the student provided.
Even if I know they have waived their FERPA rights, I share a draft of my recommendation with the student to ensure I haven’t made any mistakes (such as improper naming or gender-appropriate language/pronouns). Those detail errors are noticed by application reviewers and can undermine the candidate’s standing. The student is also assured that their journey to the profession is accurately documented.
The Office of Medical Education at the University of California San Francisco provides guidance for clinical faculty when documenting student performance, including examples of strong recommendations.9
Letterhead and signature policies
In the past, many students asked former professors to write letters of recommendation, but the schools contacted the students because those letters were not submitted on officially branded letterhead. At my institution, letterhead is restricted to official business for current employees. For those leaving the institution, I advise my peers on our letterhead policy.
Many graduate programs are more accustomed to “digital signatures” that we affix to our letters. A few of us still keep a copy of our signature or have created an artistic “signature logo” to paste into our letters. The rest just type /* their name */ (within slashes and asterisks). At any rate, admissions professionals want to see a signature, and it doesn’t take much effort to satisfy this expectation.
Final takeaways
I love learning about these students and their dreams through this process. I am reminded how inspirational these students are in the way they speak about their futures, even as I have to set aside about two to three hours per letter. I hope this guide gives you an opportunity to see how you have guided them towards their dream professions.
Checklist for Recommenders (and Reviewers)
I make sure all of these items are fulfilled when I write a reference letter.
Logistics (Recommenders only)
Timeline/Deadline
- Allow one month for scholarship/summer programs. Requests begin after fall/winter exams with deadlines in February/March.
- Allow two months for professional school applications. Requests may begin after spring midterms. Writing letters should occur after spring finals (May/June).
Conferencing Times
- Meet at least twice during your month(s) of preparation.
- FERPA waiver (signed document required from registrar and/or application service/graduate program form)
- Technology (instructions to upload, add more time for unfamiliar portals)
- Confirm that the student has set up letter receipt into the application portal.
In your letter, show your credentials
- Explain any letterhead preferences/policies
- Apply a signature
State how you know the student (drop-down questions in AADSAS)
- Role (I am the student’s professor.)
- Length (I have taught the student for Organic 1 and 2.)
- Relationship (We have not met outside of class before.)
Who needs the recommendation (i.e., Who is the Audience)?
- “Dear Selection Committee” is a typical greeting.
- Refer to the internship, scholarship, or job description.
- Leverage insights from the application drafts the student gave you.
My observations about the student (refer to the rubric)
- What stories highlight the candidate’s excellence?
- What context has the student provided that shapes your impressions?
- Which responsibilities do you trust the candidate with?
- How does the candidate stand out among other students you have taught, supervised, or advised?
- How will this candidate inspire the audience/community? How will the program help the candidate?
Checklist: Using AI Responsibly in Letters of Recommendation
Here are suggestions for maintaining your voice as an author and recommender without allowing genAI tools to dilute your evaluation or support.
- I formed my own independent evaluation of the student.
- AI was used only for organization/clarity/polish, not to invent or exaggerate content.
- I protected student confidentiality and followed institutional policy.
- The letter includes specific examples I can personally attest to.
- The voice and judgments in the letter are truly mine, and I stand by every word.
Acknowledgments
This article was peer-reviewed by and includes suggestions by Nancy Lorenzon, Teaching Professor, Department of Biological Sciences, Director of Pre-Health Education and Advising, University of Denver.
Heather Brod’s Writing Effective External Evaluation Letters inspired this article.10
Conflict of interest disclosure
No conflict of interest reported.
AI use
Generative AI tools were used to support this manuscript in the following ways:
- Documenting and summarizing private conference notes and proceedings from the 2026 Conference of the National Association for Advisors of the Health Professions (Otter.ai).
- Drafting and suggesting checklists to aid authors of letters of recommendation (Otter.ai, Gemini in-browser).
- Spelling, grammar, punctuation, and wordsmithing (Grammarly and Microsoft Copilot).
- Implementing the standard formatting for publication (ChatGPT).
References
- Guidelines for Writing a Letter of Evaluation for a Medical School Applicant, Students & Residents, AAMC website. This longstanding resource provides some tools to help authors write effective evaluation letters for applicants to medical school and residency.
- Thacker, N. (2026, revised June 19). Can You Use ChatGPT for Your Medical School Application? AMCAS, AACOMAS, CASPA, and TMDSAS Policies Compared. GradPilot. Can You Use ChatGPT for Your Medical School Application? AMCAS, AACOMAS, CASPA, and TMDSAS Policies Compared. Accessed June 30, 2026.
- Thacker, N. (2026, April 13). PA program AI policies — 10 programs with real rules. GradPilot. PA Program AI Policies — 10 Programs With Real Rules (2026). Accessed June 30, 2026.
- Coraor Fried, J., Johnson, N.R., Pelletier, A., Landman, A., Bartz, D. (2025). Using Generative Artificial Intelligence When Writing Letters of Recommendation, Academic Medicine 100 (7): 769–775. https://doi.org/10.1097/ACM.0000000000006047. Accessed June 30, 2026.
- AADSAS Letters of Evaluation FAQs (2026-2027 changes), ADEA. Also, “Preparing for ADEA AADSAS 2026” webinar, ADEA, April 21, 2026. It describes the additional characteristics questions that evaluators are asked to address.
- Chuck, E.T. How to Assess Your Academic Competencies, Student Doctor Network Review. This includes an example academic competency rubric following AAMC definitions.
- FERPA, Protecting Student Privacy, US Department of Education.
- FERPA Release for Letters of Recommendation & References at Towson University. This guidance helps faculty navigate requests for letters of recommendation.
- Writing High-Quality Evaluations of Student Performance:: Best Practices and Examples (PDF), Office of Medical Education, University of California San Francisco. Guidance for clinical faculty when documenting student performance for clerkship evaluations. It includes examples of strong recommendations.
- Brod, H. Writing Effective External Evaluation Letters, LinkedIn (2021). The inspiration for this article.
Continue reading...
