Nuance and Noise: Lessons Learned From Longitudinal Aggregated Assessment Data
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
BACKGROUND: Competency-based medical education requires frequent assessment to tailor learning experiences to the needs of trainees. In 2012, we implemented the McMaster Modular Assessment Program, which captures shift-based assessments of resident global performance. OBJECTIVE: We described patterns (ie, trends and sources of variance) in aggregated workplace-based assessment data. METHODS: Emergency medicine residents and faculty members from 3 Canadian university-affiliated, urban, tertiary care teaching hospitals participated in this study. During each shift, supervising physicians rated residents' performance using a behaviorally anchored scale that hinged on endorsements for progression. We used a multilevel regression model to examine the relationship between global rating scores and time, adjusting for data clustering by resident and rater. RESULTS: We analyzed data from 23 second-year residents between July 2012 and June 2015, which yielded 1498 unique ratings (65 ± 18.5 per resident) from 82 raters. The model estimated an average score of 5.7 ± 0.6 at baseline, with an increase of 0.005 ± 0.01 for each additional assessment. There was significant variation among residents' starting score (y-intercept) and trajectory (slope). CONCLUSIONS: Our model suggests that residents begin at different points and progress at different rates. Meta-raters such as program directors and Clinical Competency Committee members should bear in mind that progression may take time and learning trajectories will be nuanced. Individuals involved in ratings should be aware of sources of noise in the system, including the raters themselves.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.002 | 0.015 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it