Analyze student evaluations to improve your university teaching

Context

A university lecturer or professor who has just received end-of-term student evaluation results — from your institution's standard teaching survey or equivalent — and needs to turn that data into a concrete improvement plan. Reports typically arrive with item-level scores, dimension averages, and free-text comments: reading everything, identifying what to change, and prioritizing takes several hours and often stalls under the sheer volume. This prompt turns that exercise into a structured 20-minute session that ends with an actionable plan for the next term.

This is not intended for external teaching review processes (teaching portfolios, recorded classes, academic promotion) or institutional accreditation. It is personal reflection.

Originally written for Chilean classrooms. References to the CED (Cuestionario de Evaluación Docente) map to your institution's equivalent end-of-term student teaching survey.

Paste first

Before opening the model, gather:

• Your survey report: average scores by item or by dimension (usually in PDF or table format; copy as plain text)
• The response rate (number of respondents over total enrolled)
• Free-text comments from the survey, only if you can genuinely anonymize them (read the privacy note before pasting any text)
• Course name, program, term, and institution

Privacy rule — read this before pasting comments:

Student survey comments can identify individuals, especially in small groups. Before pasting:

• Remove comments that mention proper names, specific assessments known to everyone, or details that could identify a person.
• If the group has fewer than 10 responses, skip the comments and paste only the quantitative scores.
• Never paste a named report (one listing each respondent by name).
• Average scores per item are not personal data and can be pasted without restriction.

This guidance aligns with data protection regulations in most jurisdictions; apply the requirements that govern your institution.

Act as a university teaching consultant experienced in interpreting student satisfaction surveys across higher education institutions. Your role is to help me extract concrete pedagogical insights from my end-of-term student evaluation results and translate them into a realistic improvement plan for next term.

**Course context:**
- Course name: {{course_name}}
- Program or degree: {{program_or_degree}}
- Term evaluated: {{term_and_year}}
- Institution: {{institution}}
- Course profile: {{course_profile}} (total group size, response rate, delivery mode — in-person / hybrid / online, course type — lecture / practical / workshop / lab)

**Quantitative survey results (average scores by item or by dimension):**
---
{{quantitative_results}}
---

**Anonymized qualitative comments (omit this section if you cannot safely anonymize them):**
---
{{anonymized_qualitative_comments}}
---

I need you to respond in four parts:

**Part 1 — Reading the overall picture:** Which dimensions are clearly well rated? Which ones concentrate the lowest scores? If there is a notable gap between items within the same dimension, flag it. Do not repeat the numbers: interpret them.

**Part 2 — Pedagogical hypotheses:** Based on the pattern of scores (and comments, if I pasted them), formulate two or three hypotheses about what might be happening in the course that explains those results. A hypothesis is a possible cause, not a certainty. Frame them in tentative language ("this could indicate that…", "this suggests that…").

**Part 3 — Improvement plan for next term:** Propose three to five concrete pedagogical actions, ordered by expected impact. For each action, specify: (a) exactly what change it involves, (b) at what point in the term or class it is implemented, and (c) how it might show up in the next survey round if it works.

**Part 4 — One action for this week:** If the term has not yet ended or is about to begin, identify one low-cost action I can implement before the current period closes, without waiting for the full redesign.

Do not soften the findings. If an area is systematically poorly rated, tell me directly. Use the language of a professional peer, not an external evaluator.

Expected output

Example (condensed):

Part 1 — Reading the overall picture

Availability (4.5) and syllabus adherence (4.3) are the pillars of the course.
But there is a systematically low cluster: clarity of explanations (3.2),
active participation (3.1), and alignment between assessments and content (3.4).
The 1.4-point gap between availability and active participation is the most
important signal: the one-on-one relationship works; the shared classroom
space does not.

Hypothesis 2: There is a mismatch between the cognitive level of in-class
practice and the assessments — exercises calibrated toward direct application,
assessments requiring transfer. That leap is not being scaffolded.

Plan — priority action: In at least one lecture per unit, interrupt the
exposition with a short pair problem (5–10 min) before the whole-group
discussion. It moves participation and clarity simultaneously.

This week: Distribute two or three "assessment-level" problems before the
final evaluation. Cost: two hours. Closes the gap before the term ends.

Watch out for

• Student surveys measure perceived experience, not actual learning. A low score on "clarity of explanations" may mean your explanations are unclear — but it may also mean the content is genuinely difficult or that the group expected less rigor. The model will work with the most pedagogical hypothesis; you have the context to validate or discard it.
• If you paste only quantitative scores without comments, the hypotheses will be more general. That is preferable to pasting comments that could identify a student.
• The model may propose changes that require institutional infrastructure (online response systems, co-teaching, specialized rooms). If that happens, ask for alternatives that fit within your current resources.

Suggested iteration

If the hypotheses feel too generic: "Hypothesis 2 could apply to any university course. Go back to the item with the lowest score and tell me what specific elements of the context I described could explain that result in particular."

If you want to adapt the plan to your classroom reality: "I have a room of 80 students, no teaching assistants, and two hours per week. Of the proposed actions, which are feasible in that context? Adjust the plan."

If you need to report the results to your department head or teaching coordinator: "Write a summary paragraph of the survey results that I can use in my end-of-term management report, in an objective tone and without revealing the raw scores."