School climate and SELAssessment and academic integrityLesson and material designPrivacy and student data

Diagnose school climate from behavior and wellbeing data

Tested on
Claude Sonnet 4.6, noviembre 2026
Estimated time
30 min
Time saved
3-4 hours
Published
2026-11-16
Last reviewed
2026-11-16
Attribution
Equipo Circles

Context

A school principal or head of curriculum who has access to behavior and wellbeing data from the semester or year — incident reports, behavior log entries, school climate surveys — and needs to turn that data into a useful diagnosis before reviewing or updating the school's wellbeing plan. This prompt covers the analysis step that comes before designing interventions: detecting patterns (by grade band, level, gender, incident type, and period), identifying the most critical risk factors, and generating a prioritized action plan. This is the step that comes before drafting or updating a code of conduct; without this diagnosis, any new protocol starts from assumptions rather than institutional evidence.

Originally written for Chilean classrooms. References to "MBE" (Marco para la Buena Enseñanza) are adapted here as "your school's teaching framework"; "apoderados" as "parents/guardians"; and "Bases Curriculares" as "your national curriculum". The Chilean legal reference (Ley 19.628) addresses personal data protection — apply your jurisdiction's equivalent student privacy regulation.

Best timing: end of semester or year, before the faculty evaluation meeting and before starting the review of the school's behavior and wellbeing code.

Paste first

Before opening the model, have ready the following inputs. All must be anonymized:

  • School description: type (public, charter, independent) and the grade levels served (elementary, middle, high school, or combined)
  • Incident data: total behavior reports and wellbeing log entries, broken down by incident severity (minor, serious, severe), grade band or level, month or period, and — if available — gender breakdown as a percentage of the group (never as individual counts)
  • School climate survey results (if conducted): scores or percentages by dimension — peer relationships, student-teacher relationship, perception of safety — broken down by grade band or level where possible
  • Relevant institutional context: Is there a school improvement plan focused on wellbeing? A dedicated wellbeing coordinator or counselor? Any significant event during the year that might explain variations (schedule change, construction, external events)?

Privacy rule: Never paste national ID numbers, student names, parent/guardian names, or staff names. Do not include individual referral forms, individual psychoeducational reports, or incident reports with personal identifiers. Always work with totals and percentages by group, grade band, or level. This applies even if the model is private or running locally — apply your jurisdiction's student data privacy regulation.

Prompt

Act as a school wellbeing specialist and institutional data analyst. I am going to share aggregated incident data and climate survey results from my school, and I need your help building a wellbeing diagnosis and a prioritized action plan.

**School information:**
- School type: {{school_type, e.g. "public charter school serving grades K-12"}}
- Grade levels served: {{levels, e.g. "kindergarten through grade 12"}}
- Number of students: {{number_of_students}}
- Period analyzed: {{period, e.g. "first semester 2026"}}
- Relevant context: {{context, e.g. "school improvement plan with wellbeing focus in its second year of implementation, full-time wellbeing coordinator, school counselor two days per week"}}

**Incident data (anonymized — no names or ID numbers):**
{{Paste here the breakdown of behavior reports and wellbeing log entries. This can be a table copied from a spreadsheet, free text, or a list. Include: total by incident severity (minor/serious/severe), breakdown by grade band or level and by month or quarter. If available, include gender breakdown as a percentage of the group, not as an individual count.}}

**School climate survey results (if available):**
{{Paste here the results of the most recent climate survey, with scores or percentages by dimension and by grade band. If no survey was conducted, write "not available" and the model will adjust its analysis.}}

**Comparison with previous period (optional):**
{{If you have data from the same period in the previous year, paste it here to analyze trends. If not available, write "not available".}}

With this information, please generate:

1. **Pattern diagnosis:** identify the 3 to 5 most relevant patterns in the data, specifying which grade band, level, incident type, or period concentrates the highest incidence. For each pattern, formulate an explanatory hypothesis the team can discuss, distinguishing clearly between correlation and causality.

2. **Risk factor map:** from the patterns identified, name the institutional factors most likely at play (classroom climate, recess supervision, transition periods, relational dynamics). Flag which factors are supported by the data and which are hypotheses to validate with the team.

3. **Prioritized action plan:** propose 4 to 6 concrete measures, ordered by potential impact and feasibility for the following semester. For each measure, state: what problem it addresses, who is the main responsible party, what timeline is realistic, and how its effect would be evaluated.

4. **Two generative questions** to open discussion at the faculty meeting or with the leadership team: one that challenges a hypothesis in the diagnosis and one that invites reflection on the resources already available at your school.

Do not invent data or assign causality that is not supported by the numbers I have provided. If the information is insufficient for any point, mark it with [NEEDED: what data is missing] rather than inventing it.

Expected output

Example (trimmed):

Pattern diagnosis:

Pattern 1: Grades 5–8 account for 44% of behavior reports (+48% year-over-year) and show the lowest safety perception scores in the school (2.9/5). Hypothesis: insufficient adult supervision during recess and transitions. [NEEDED: location-of-incident records to validate]

Pattern 2: Grades 9–10 grew +59%, with a peak in April coinciding with an extended school day change. Hypothesis: schedule reconfiguration disrupted relational dynamics.

Pattern 3: 64% of teachers perceive inconsistent application of the school's behavior code; "clear and consistently applied rules" scores lowest (2.7). This is the factor with the greatest cross-cutting presence in the data.

Action plan (top 3):

  1. Unified discipline protocol (July–August). Goal: reduce from 64% to ≤40% of teachers perceiving inconsistency.
  2. Targeted intervention for grades 5–8: social-emotional skills sessions and reorganization of recess supervision rotations. Goal: reduce reports from 65 to <50.
  3. Support program for grades 9–10 after the extended school day change, with a monthly dialogue session with class representatives.

Generative question: Is the inconsistency a matter of individual teacher judgment, or does the behavior code have grey areas that make uniform application impossible?

Watch out for

  • Personal data: behavior reports and wellbeing log entries are linked to identifiable students. Before copying any table, remove ID numbers, names, file numbers, and individual diagnoses. The model can generate a useful diagnosis from totals and percentages without needing individual data — apply your jurisdiction's student privacy regulation.
  • Causality vs. correlation: the model can propose plausible explanatory hypotheses, but patterns in data do not prove causes. Every finding should be treated as a starting point for team discussion, not as a validated conclusion. This is especially important when the model associates incidents with group characteristics — grade band, gender — without recognizing the contextual factors your school knows and the model does not.
  • Measures disconnected from your school's reality: the action plan may propose measures that already exist at your school or that are not feasible given your current budget or staffing. Review every measure against your real context before presenting it to the leadership team or faculty meeting.

Suggested iteration

If the diagnosis is too generic: "We already knew about pattern 2 (increase in conflicts in grades 7 and 8). Go deeper on that level: what type of incident predominates, at what time of day does it appear most, and which hypotheses are consistent with the data I provided?"

If you want to prepare the presentation for the faculty meeting: "Turn the diagnosis and action plan into a four-slide structure: key findings, risk factor map, prioritized plan, and questions for the team. Keep the same content and adapt only the format."

If you need to communicate results to families: "Write a paragraph of no more than five lines to share the wellbeing findings with parents/guardians at the end-of-semester meeting, in plain language and without exposing data that could identify specific groups of students."