SLO PI rationale methodology
Purpose:
This standalone page explains workbook PI scores using visible syllabus/course-context evidence. The output is intended to help course admins understand whether the score is easy to defend, what evidence is missing, and what concrete course artifact could make the alignment clearer.
Core prompt:
Analyze the supplied course syllabus/course context against the SLO/PI rubric and workbook PI scores. Preserve workbook scores unless explicitly asked to rescore. For each PI, classify the visible evidence as direct evidence, direct but incomplete, inferred evidence, or no visible evidence. Produce Rationale and Advice by default, plus Example only when the score is below 3 or a score of 3 is not fully defensible from visible evidence. Be constructive: distinguish course quality from accreditation-visible evidence, and when the course may already do the work but the syllabus does not show it, say that the fix may be documentation rather than course redesign.
Default output fields:
- Rationale: why the workbook score is defensible, limited, or hard to defend from visible evidence.
- Advice: what should be added or clarified to improve a low score or defend a high workbook score.
- Example: a course-specific artifact, wording, task, assessment, report section, lab requirement, design review, or reflection. Omit this when a score of 3 is fully defensible from direct evidence.
Verbose-only field:
- Keep explicit: what evidence should remain visible for future reviewers and where citation hover cards should point.
Evidence policy:
- Direct evidence must show students doing the PI action in an outcome, task, assessment, project, lab, report, or rubric.
- Direct but incomplete evidence shows a relevant action, but misses an important PI clause, constraint, context, or assessed deliverable.
- Inferred evidence can support cautious wording, but should not be described as direct PI support.
- No visible evidence means the supplied materials do not yet show the PI-level action.
- Do not infer ethics, sustainability, communication, teamwork, lifelong learning, or societal impact from technical work alone.
- If a high workbook score has only nearby evidence, write what other information would defend the workbook score.
- If a low score has related technical content, name the missing PI action instead of criticizing the course.
Constructive example pattern:
Advice: Add a task where students interpret the consequences of a computational result, not only whether the computation is correct.
Example: In the final Python modelling report, add a short "model limits and decision impact" section where students explain one assumption, one source of numerical error, and one practical consequence of acting on the result.
Course-specific artifact examples:
- 30.001 Structures & Materials: structural design calculation report.
- 30.002 Circuits & Electronics: circuit lab/design report.
- 30.007 Engineering Design Innovation: prototype design review.
- 30.100 Computational and Data-Driven Engineering: Python modelling report.
- 30.101 Systems and Control: controller design report.
- 30.102 Electromagnetics & Applications: EM modelling or application project brief.
- 30.103 Fluid Mechanics: fluid analysis project report.
QA checklist:
- Every PI card shows Rationale and Advice by default.
- Example appears only for score < 3 or score-3 PIs that are not fully defensible from direct evidence.
- Every PI card has one evidence label: direct evidence, direct but incomplete, inferred evidence, or no visible evidence.
- Direct and direct-but-incomplete claims carry hover-ready citation metadata or a placeholder.
- Keep explicit appears only when verbose mode is on.
- Rationale does not overclaim direct evidence from weak topic overlap.
- Advice is constructive and says what to make visible.
- Example names a plausible artifact from the course.
- Scores, totals, workbook source, selections, and profiles are not changed by rationale text.