C-OBE cheat sheet

Outcome-Based Education organizes curriculum, delivery, and assessment around the culminating outcomes learners must demonstrate, rather than around inputs such as content coverage or credit hours.

Spady's four principles of OBE are clarity of focus, design down, high expectations, and expanded opportunity.

Design down means designing backwards from exit outcomes to program outcomes, then to course outcomes, then to unit-level objectives.

Spady distinguishes traditional OBE (flexible use of existing structures), transitional OBE (outcomes describing existing curricula), and transformational OBE (future-oriented role outcomes).

Revised Bloom's Taxonomy (Anderson & Krathwohl, 2001) orders the cognitive process dimension as Remember, Understand, Apply, Analyze, Evaluate, Create, crossed with factual, conceptual, procedural, and metacognitive knowledge.

Miller's pyramid climbs from Knows to Knows how to Shows how to Does; only 'Shows how' and 'Does' are accepted as performance- or workplace-based evidence.

Mastery learning (Bloom, 1968) fixes the performance standard and varies time, using formative checks with corrective instruction until the criterion is met.

Biggs's 3P model separates presage (student factors and learning context), process (learning activities), and product (outcomes), so outcomes are treated as the product of alignment rather than of inputs alone.

Program educational objectives are broad, career-oriented statements describing performance three to five years after graduation; program learning outcomes describe performance at graduation and are directly measured.

ABCD template: Audience (who), Behavior (observable verb), Condition (context and tools allowed), Degree (criterion standard, e.g., 80% accuracy).

Each outcome carries a single observable verb; verbs such as know, understand, appreciate, and be aware are not observable and must be replaced with identify, explain, apply, analyze, or design.

SMART test for an outcome: specific, measurable, achievable, relevant, time-bound.

The cognitive level of a course outcome must be at or above the level of the program outcome it supports; a PLO requiring 'design' cannot be served only by CLOs at 'list' or 'describe'.

Every program outcome must be mapped to at least one course, and preferably to several, so that attainment rests on more than a single evidence source.

The Washington Accord defines 12 graduate attributes: engineering knowledge, problem analysis, design/development of solutions, investigation, modern tool usage, engineering and society, environment and sustainability, ethics, individual and teamwork, communication, project management and finance, and lifelong learning.

Course outcomes are stated from the learner's perspective ('students will be able to...'), not from the teacher's ('this course covers...').

Constructive alignment (Biggs) requires three components to target the same verb and level: intended learning outcomes, teaching and learning activities, and assessment tasks.

Backward design (Wiggins & McTighe) has three stages: identify desired results, determine acceptable evidence, then plan learning experiences and instruction.

Decision rule: the verb in an assessment task must match or exceed the verb in the aligned outcome; assessing 'create' with recall items is a misalignment.

SOLO levels are prestructural, unistructural, multistructural, relational, and extended abstract; multistructural responses list ideas while relational responses integrate them.

An alignment matrix places outcomes as rows and activities or assessments as columns; empty cells identify outcomes with no instruction or no assessment.

Formative tasks should rehearse the exact performance the summative task requires, so feedback transfers directly to the graded evidence.

Surface learning reproduces content for the test, whereas deep learning requires activities and assessments that demand integration and application to unfamiliar contexts.

Assessment is the process of gathering evidence of learning; measurement assigns a quantitative value to that evidence; evaluation judges worth or quality against criteria.

Criterion-referenced assessment compares performance to a fixed standard; norm-referenced assessment ranks performance against other examinees.

Validity types: content (alignment with the table of specifications), construct (measures the intended ability), and criterion (correlates with an external benchmark).

Reliability types: internal consistency (Cronbach's alpha, 0.70 or above acceptable), test-retest (stability over time), inter-rater (agreement between scorers), and split-half.

Item difficulty p = (number answering correctly) divided by (number of examinees); p between 0.30 and 0.70 discriminates best, and p above 0.90 signals an item that is too easy.

Discrimination index D = p(upper 27%) minus p(lower 27%); D of 0.20 or more is usable, D of 0.40 or more is good, and negative D flags a defective item to revise or discard.

A table of specifications sets item counts per topic and cognitive level: items for a topic = (hours on topic divided by total hours) times total items.

Kirkpatrick's four evaluation levels are reaction, learning, behavior, and results; end-of-course satisfaction surveys measure only the first.

A curriculum map is a matrix of courses (rows) against program outcomes (columns) coded I, R, M, A: Introduced, Reinforced, Mastered, Assessed.

A program outcome whose map shows only I entries, with no M or A, is taught but never brought to mastery or measured.

Vertical alignment sequences outcomes across year levels; horizontal alignment coordinates outcomes across courses delivered in the same term.

The syllabus documents course outcomes, their mapping to program outcomes, assessment tasks with weights, the schedule, and learning resources; it is the public delivery contract.

Learner support covers academic advising, remediation and tutoring, mentoring, and access services; these are inputs that must be evidenced, not assumed, in accreditation.

The continuous improvement delivery loop runs: map outcomes, deliver instruction, collect assessment data, analyze gaps, implement change, reassess.

Active-learning delivery requires students to produce observable work in class (solve, discuss, build, present), which supplies formative evidence for alignment checks.

PDCA (Plan-Do-Check-Act) is the standard continuous-improvement cycle; each turn records the plan, the action taken, and the verified result.

ABET's eight general criteria are students, program educational objectives, student outcomes, continuous improvement, curriculum, faculty, facilities, and institutional support.

Closing the loop requires documented evidence of three steps: results analyzed, an action implemented, and later data showing the effect of that action.

A self-assessment report states, for each criterion, what the program does, what evidence demonstrates it, and what improvement is planned.

The Washington Accord recognizes substantial equivalence of accredited engineering programs among signatories and is the reference for the 12 graduate attributes.

Accreditation evidence must be traceable: every claim points to a specific artifact such as an assessment report, rubric, meeting minutes, or sample of student work.

Stakeholder input from students, alumni, employers, and industry advisory boards feeds both program educational objective review and curriculum improvement decisions.