What Surgeons-in-the-Making Reveal: Decoding the Personal Statement, One Code at a Time

Every year, roughly 5,500 applicants vie for a fixed number of categorical positions in general surgery residency programs across the United States, and a single program can receive around 1,100 applications in one cycle. Screening that mountain of paperwork in a matter of weeks has long pushed programs toward blunt instruments: USMLE score cutoffs, letters of recommendation, and membership in honorary societies such as Alpha Omega Alpha. But the landscape is shifting. With Step 1 now reported as pass or fail, the traditional numerical anchors of the application have lost much of their discriminating power, and programs are being forced to ask a harder question: what actually predicts a successful surgeon? A new study from Washington University in St. Louis offers an unusually detailed answer by taking a scalpel to one of the most subjective documents in medicine, the personal statement.

The research, published in Global Surgical Education, the Journal of the Association for Surgical Education, is a retrospective qualitative content analysis of personal statements submitted during the 2023-2024 application cycle to the categorical general surgery residency program at WashU Medicine. Led by Usman Y. Panni and Tiffany K. Brocke, the team drew a stratified random sample of 294 applicants from the 620 who applied from LCME-accredited medical schools, preserving the demographic distribution of the full pool by stratifying on self-designated gender and status as underrepresented in medicine, or URiM. The goal was not to grade the essays but to map what applicants actually write about, and whether those themes differ systematically across groups.

The methodological machinery is where the study gets technically interesting. Two lead coders developed a codebook inductively by reviewing previously available residency personal statements, reconciling overlapping concepts until they arrived at 24 distinct codes organized into six categories: humanism, teamwork, expertise, discipline, discovery, and adaptability. Each code came with an operational definition and illustrative examples, and a code was flagged as strongly emphasized when the applicant developed it through a substantive personal experience tied to their path into medicine or their professional goals. Before full coding, the leads jointly reviewed 25 statements to calibrate their approach, then trained teams of reviewers, eight residents at the PGY3 or PGY4 level and two medical students, who each independently coded every statement, with disagreements resolved by discussion or escalation to the leads.

To strip out biasing cues, a senior program coordinator redacted every statement before coding, removing names, gendered pronouns, geographic locations, and institutional identifiers. The de-identified documents were then imported into ATLAS.ti, a qualitative analysis platform, for systematic coding. Demographic and academic variables, including age, gender, AOA status, school type, and interview invitation status, were extracted from ERAS applications and linked to the coded content through anonymized alphanumeric identifiers. The analysis itself, run in RStudio and SPSS, used chi-square and Fisher’s exact tests to probe associations between coded content and applicant characteristics. Notably, the content analysis was performed retrospectively, after the Match, so it played no role in who was actually invited to interview.

The headline finding is that personal statements are far more structured than their free-flowing prose suggests. On average, each statement expressed ten of the 24 codes, with a range from 3 to 16. At the category level, humanism and discovery dominated, reflecting applicants’ emphasis on patient-centered values, personal growth, and intellectual curiosity. The five most common individual codes were self-discovery, present in 78.2 percent of statements, mastery at 67.0 percent, working in teams at 61.5 percent, compassion at 60.2 percent, and responsibility at 55.1 percent. Self-discovery captured moments when applicants connected past experiences to their decision to pursue surgery; mastery marked descriptions of gaining technical or nontechnical proficiency; compassion and responsibility reflected perceptions of the obligations inherent in surgical practice.

At the other end of the spectrum, some professional values barely registered. Leadership in the context of power appeared in only 18.3 percent of statements, followed by supportive community at 23.8 percent, flexibility at 25.8 percent, cultural competence at 25.8 percent, and strategy at 26.2 percent. The pattern held for strong emphasis as well: when applicants backed a code with a detailed personal narrative, mastery led at 19.4 percent, followed by individual patient connections at 13.6 percent and compassion at 12.2 percent, while strategy, supportive community, and cultural competence were strongly emphasized in fewer than 3 percent of statements. In other words, the qualities applicants write about most are not always the ones they can substantiate with concrete experience, and some values central to surgical practice may be systematically underrepresented in applicant narratives.

The demographic breakdowns are where the study becomes genuinely provocative. Applicants who identified as men were significantly more likely than women to emphasize the application of science to patient care, at 51.2 percent versus 37.0 percent. URiM applicants were markedly more likely to highlight justice, at 55.6 percent versus 27.1 percent, cultural competence at 51.9 percent versus 19.6 percent, leadership framed as power at 29.6 percent versus 15.8 percent, and resilience at 66.7 percent versus 45.0 percent. Conversely, URiM applicants less frequently emphasized the application of science to patient care, at 24.1 percent versus 47.5 percent, and problem-solving, at 27.8 percent versus 46.2 percent. The authors interpret these differences as reflections of divergent lived experiences, mentorship exposure, and perceptions of what programs value, and they warn that equivalent qualities may be expressed in different narrative registers depending on a writer’s background.

The comparison with interview outcomes adds a layer of real-world consequence. Among the 94 sampled applicants invited to interview, intellectual curiosity appeared in 60.6 percent of statements versus 44.0 percent for uninvited applicants, and application of science to patient care in 57.4 percent versus 36.0 percent. Uninvited applicants, by contrast, more often emphasized communication, at 36.5 percent versus 21.3 percent, and cultural competence, at 30.0 percent versus 16.0 percent. In strength-of-emphasis analyses, invited applicants more often strongly emphasized justice and responsibility. Intriguingly, mean USMLE Step 2 CK scores did not differ significantly between invited and uninvited applicants, at 254.4 versus 251.0, suggesting that the narrative content of the statement was picking up signal that the numbers were not. The authors are careful to stress that these are associations, not causal claims, since the coding was retrospective and other application materials also informed interview decisions.

Placed alongside prior work, the study occupies a distinctive methodological niche. Ostapenko and colleagues used software-driven textual analysis to identify gender-associated patterns in tone and topic; Naples and colleagues found that highly ranked applicants used more grit-related language, particularly women; and Stain and colleagues showed that personal statements rated as excellent were associated with higher applicant rankings, on par with Step scores and AOA membership. The WashU team instead performed a calibrated, human review mimicking the faculty review that happens every Match cycle, applying a shared institutional codebook with consensus checks. That design, they argue, makes the approach portable: programs could adopt a codebook, train reviewers, and standardize how narrative information enters holistic review, reducing reliance on global impressions and guarding against over- or under-weighting codes based on a writer’s identity or narrative style.

The limitations are candidly acknowledged. The codebook was internally developed and may not generalize to other programs or applicant populations; manual coding remains inherently subjective even with de-identification and consensus adjudication; individual pre-consensus ratings were not retained, so formal inter-rater reliability could not be calculated; and no a priori power analysis was performed for this exploratory study. The retrospective, cross-sectional design limits causal inference, and the codebook may not capture all relevant content. Still, the implications are hard to ignore. If structured content analysis can make explicit which experiences and attributes reviewers are recognizing, it could add a standardized qualitative dimension to resident selection, inform faculty reviewer training, and, with longitudinal follow-up and multi-institutional validation, ultimately reveal whether the values applicants write about today predict the surgeons they become tomorrow.

Subject of Research: Content analysis of personal statements in general surgery residency applications and its implications for holistic applicant review

Article Title: Content analysis of general surgery residency personal statements: implications for holistic review

Article References: Content analysis of general surgery residency personal statements: implications for holistic review. (n.d.). https://doi.org/10.1007/s44186-026-00605-4

Image Credits: AI Generated

DOI: 10.1007/s44186-026-00605-4

Keywords: general surgery residency, personal statements, holistic review, content analysis, medical education, URiM applicants, residency selection, USMLE, qualitative coding, interview invitation, gender differences, surgical training