Guide · Methodology

Understanding FCQ Scores at CU Boulder

Every subscore on the University of Colorado Boulder Faculty Course Questionnaire, what it actually measures, and how DegreeDraft turns raw FCQ data into the composite rating you see on every professor page.

What the FCQ is

The Faculty Course Questionnaire is an end-of-term survey every CU Boulder student is asked to complete for each course they take. Responses are aggregated per instructor and per course section and published by the CU Boulder FCQ office. Because the survey is gated to enrolled students and every response is tied to a specific course-section, it is the most rigorous rating dataset available for CU Boulder — far more reliable than public review sites that anyone can post to.

The four subscores DegreeDraft shows

Each FCQ question is reported on a 1–5 scale. DegreeDraft surfaces four rollups:

Teaching effectiveness — how well the instructor explained material, paced class, and responded to questions.
Feedback quality — usefulness and timeliness of feedback on exams, papers, and assignments.
Instructor availability — how accessible the instructor was outside class (office hours, email, post-lecture conversation).
Course difficulty — how challenging students found the material and the workload.

On professor pages we show the raw subscore averages alongside the sample size (the number of FCQ responses). On course pages we aggregate difficulty across every instructor who has taught the course in the indexing window, so the difficulty signal survives instructor turnover.

Why we use Bayesian smoothing

A raw average over 8 responses is noisy — one outlier can swing a score by 0.3 points. To avoid unfairly ranking a low-sample course against a high-sample one, DegreeDraft applies Bayesian smoothing: the observed average is pulled toward the campus-wide mean, with the strength of the pull proportional to the inverse of the sample size.

Practically: a course with 10 responses and a 4.8 raw average gets smoothed down toward (say) 4.3. A course with 400 responses and the same 4.8 raw average barely moves. The result is a ranking that prefers courses with real evidence of being well-taught over courses that happened to draw a handful of enthusiastic reviewers.

We only display smoothed scores once a course crosses the 10-response threshold. Below that, the course is hidden from ranked guides like easiest courses to avoid headline numbers that could mislead.

How the composite rating blends FCQ and RateMyProfessor

When DegreeDraft has both FCQ and RateMyProfessor data for an instructor, the composite score blends them with inverse-variance weighting — each source contributes in proportion to how many responses it has. When sample sizes are comparable, FCQ carries more weight because it is gated to CU Boulder enrollees, while RateMyProfessor is open to anyone.

RateMyProfessor also has known self-selection bias: students who feel strongly (positive or negative) are much likelier to post. DegreeDraft flags RMP entries with fewer than 5 ratings as low-confidence and de-emphasizes them in the composite. The "would take again" percentage is surfaced separately because it's the RMP signal that correlates best with FCQ teaching scores.

Difficulty is not the opposite of quality

A common misread of FCQ data is to assume that "hard = bad teacher." Those are independent dimensions. An organic chemistry course can be genuinely difficult — deeply abstract material, heavy homework — while being taught by an excellent instructor who scores 4.8/5 on teaching effectiveness. DegreeDraft keeps difficulty and teaching scores separate so you can make the trade-off deliberately instead of collapsing them into one "is this class good" number.

FCQ methodology FAQ

What is an FCQ at CU Boulder?

FCQ stands for Faculty Course Questionnaire — the end-of-term evaluation every CU Boulder student is invited to complete for each course. Responses are aggregated per instructor and per course and published on the CU Boulder FCQ dashboard.

What do the FCQ subscores measure?

The four subscores DegreeDraft surfaces are: teaching effectiveness (how well the instructor taught), feedback quality (quality and timeliness of feedback on student work), instructor availability (how available the instructor was outside class), and course difficulty (how challenging students found the material). All are reported on a 1–5 scale.

Why does DegreeDraft use Bayesian smoothing on FCQ scores?

Raw FCQ averages with 5 or 10 responses are noisy — one outlier can swing the score dramatically. Bayesian smoothing pulls low-sample scores toward the campus-wide mean, so a course with 8 responses isn't unfairly ranked against a course with 800. Only courses with at least 10 responses are shown; lower-sample data is hidden behind a low-confidence flag.

How does the composite rating blend FCQ and RateMyProfessor?

DegreeDraft merges FCQ teaching-effectiveness with RateMyProfessor overall-quality using inverse-variance weighting, which favors whichever source has more responses. FCQ is preferred when sample sizes are comparable because it is gated to CU Boulder-enrolled students, while RateMyProfessor is open to anyone.

Is RateMyProfessor data reliable at CU Boulder?

RateMyProfessor has self-selection bias — students who feel strongly (positively or negatively) are likelier to post. It's useful as a secondary signal, especially for "would take again" percentages, but DegreeDraft flags low-sample RMP entries (fewer than 5 ratings) as low-confidence and weights FCQ data more heavily when both are available.

Why is the difficulty score sometimes different from the FCQ teaching score?

They measure different things. An instructor can be effective (high teaching score) while the material itself is hard (high difficulty score). DegreeDraft's difficulty is a composite of FCQ challenge, grading rigor, and GPA signals — so "easy" in this context means low workload and approachable grading, not necessarily worse teaching.

FCQ data is collected and published by the CU Boulder Office of the Faculty Course Questionnaire. DegreeDraft processes the public FCQ export with Bayesian smoothing and blends it with RateMyProfessor data where available. All methodology decisions — thresholds, weighting, smoothing — are our own; FCQ and RateMyProfessor are not affiliated with DegreeDraft.