Michelle Blomberg

Dial Your Course

Case study · AI Tools & Strategy · Learning Design

A full audit of one real course package, every finding traced to its standard and its location.

Report · GC Blomberg - AVC100 Sandbox

28 findings across 34 pages, 27 graded items, and 17 rubrics
2 to fix 11 to review 15 satisfied
Every figure below was computed from the course package. A check that cannot compute does not appear, and no standard is reported as satisfied unless it was verified.

Does the course hold together?

Outcomes, evidence, and workload.

Canvas setup
pass

Outcomes · complete. 14 outcomes, 13 of 17 rubrics bound, 14 criteria carry an outcome

Canvas → Outcomes

Outcomes exist as objects and rubric criteria point at them. Alignment is evidenced rather than asserted, this report builds the matrix from the course itself, and Canvas can report outcome mastery in the gradebook.

Outcome to assessment alignment
pass

Every stated outcome is assessed

Course-wide

All 14 outcomes are carried by at least one rubric criterion on a graded item. The alignment matrix closes.

review

14 of 27 graded items carry no outcome

Course-wide

Graded, but attached to no course outcome:

Exercises and Reflection
  • Student Success: Reputable Tutorial/Learning Sources (3 pts)
  • Industry Inspiration: Graphic Design (5 pts)
  • Industry Inspiration: Photography (5 pts)
  • Student Success: Advising (3 pts)
  • Student Success: Campus Cares (3 pts)
  • Syllabus and Course Introduction (5 pts)
  • Student Success: Where to get help (3 pts)
  • Student Success: Online Learning Tips (3 pts)
  • Industry Inspiration: Animation (5 pts)
  • Course Reflection (10 pts)
  • Student Success: Digital Media Programs at the college (3 pts)
  • Introduce yourself (5 pts)

Projects
  • Photographic Postcard, Step 3: Refine (15 pts)
  • Social Media Animation, Step 3: Refine (15 pts)

This is not automatically a defect. It is a question with three answers: a missing link, deliberate co-curricular work, or busywork. The tool will not decide for you, but it requires that you decide.

What to do

Declare why, and this stops being a finding. Unaligned graded work is not automatically a defect. The standard this instrument enforces is not that every assessment map to a course outcome, but that you can say what every assessment is for. A declared exception is a design decision. An undeclared one reads as an accident, and a reviewer cannot tell the difference from the outside.

Exercises and Reflection 12 items
Projects 2 items

pass

Grade weighting is on, and the groups sum to 100%

Course settings

group_weighting_scheme is percent, so Canvas is enforcing these weights:

• Exercises and Reflection, 45%
• Projects, 55%

Confirm these match the figures published in your syllabus. The syllabus is the promise; the gradebook is what happens, and nothing in Canvas keeps them in agreement.

review

Grade weighting · reconcile the gradebook against the syllabus

Course-wide

Enforced by the gradebook (assignment group weights):
• Exercises and Reflection, 45%
• Projects, 55%

These weights sum to 100%.

The syllabus is the promise; the gradebook is what happens. Nothing in Canvas keeps them in agreement, and students discover the gap while calculating what they need on a final.

What to do
  1. Compare the two lists above line by line. Where they disagree, decide which is correct, then fix the other.
  2. Confirm every graded item sits in the group you believe it is in. An assignment in the wrong group is weighted at the wrong rate, and nothing warns you.
  3. Check for a weighted group with no assignments, and for assignments in a zero-weight group. Both are silent and both are common.
Course competencies
pass

Competency map · 14 of 14 competencies are evidenced by a rubric criterion

Rubric criteria

Each competency is listed with the assignment and the specific rubric criterion that assesses it. This is read from the binding on the criterion, not inferred from assignment titles.

CompetencyWhere it is assessed
Build and defend a colour paletteColor Palette Exercise · 2 pts
Color choices are appropriate to each client’s audience
Draw with the Pen toolLesson and Exercise: Illustrator · 5 pts
5+ pen tool shapes
Sketch and select a poster compositionGraphic Design Poster, Step 1: Sketch · 3 pts
Thumbnail sketches: 20 (poster) / 10 (postcard) / 4 to 6 storyboard frames (animation)
Build a vector travel posterGraphic Design Poster, Step 2: Build · 4 pts
Core production work: vector art (poster) / compositing (postcard) / keyframing (animation)
Revise from critique and export for printGraphic Design Poster, Step 3: Refine · 2 pts
File prepared for output: correct resolution, color mode, and size. The PDF would print. The MP4 would play. The JPG is at delivery size.
Isolate a subject in PhotoshopLesson and Exercise: Photoshop · 5 pts
Three subjects cut out from their backgrounds
Plan a postcard that extends the posterPhotographic Postcard, Step 1: Sketch · 1 pts
Typeface selected and justified (poster) or carried forward (postcard, animation)
Composite a layered postcardPhotographic Postcard, Step 2: Build · 4 pts
Animal isolated and placed cleanly, edges refined
Animate with keyframesLesson and Exercise: After Effects · 5 pts
Fade with opacity keyframes
Storyboard a motion sequenceSocial Media Animation, Step 1: Sketch · 3 pts
Thumbnail sketches: 20 (poster) / 10 (postcard) / 4 to 6 storyboard frames (animation)
Build and export an animationSocial Media Animation, Step 2: Build · 4 pts
Core production work: vector art (poster) / compositing (postcard) / keyframing (animation)
Curate and publish a portfolioBehance Portfolio · 2 pts
Project with descriptive title
Operate the course toolchainTechnology Orientation · 3 pts
Adobe CC with Illustrator, Photoshop, After Effects
Capture physical work for critiqueGraphic Design Poster, Step 1: Sketch · 2 pts
Sketches captured for critique: photographed or scanned square to the page, evenly lit, in focus, linework readable at full size in Miro. If a classmate cannot see it, they cannot critique it.

This tool cannot fetch the board record. The Curriculum Office does not serve its pages to scripts from another origin, so no browser tool can retrieve the official text automatically. Open the record for AVC100 ↗ and read the competencies against the left column.

What to do
  1. Confirm each criterion genuinely evidences the competency it is bound to. The binding is machine-readable; its correctness is not.
  2. Check the effective dates on the board record. Competencies change between approved versions.
Measurable objectivesOSCQR 9
pass

All 14 objectives use observable verbs

Learning outcomes

Each objective names a performance that can be watched and scored.

Assessment validity under generative AIOSCQR 31, 45
fix

1 graded item could be completed by a generative model

Graded work

Each item below matches a pattern that a current model satisfies in under a minute. This is a construct validity finding, not an integrity accusation. An assessment that something other than the student can satisfy has ceased to measure the student.

Course Reflection (10 pts)
A prose response to a general prompt, submitted as text or a single upload
A model produces a competent version of this from the prompt. Without process evidence, the artifact cannot distinguish the student from the model.

What to do
  1. Move the evidence from the artifact to the process. Require drafts, iteration, and a documented response to critique received. A model can produce the artifact; it cannot produce the student’s revision history in response to a named classmate’s comment.
  2. Situate the prompt. A question answerable without reference to this section, this critique, this local context, is a question answerable without the student. Bind the prompt to something that exists only in this course this term.
  3. Assess the defense, not only the product. A short recorded or live walkthrough of the decisions made converts an artifact into a performance.
  4. Recall quizzes: keep them if they are low-stakes practice, and stop treating the score as evidence of mastery. If the score carries weight, replace it with an application task at the same cognitive level.
  5. Authentic does not mean AI-resistant. A model can write a portfolio case study. The artifact never protects the assessment; the process does.
AI resistance, item by item
review

AI resistance · 9 of 27 graded items collect strong process evidence

Graded work

An artifact never protects an assessment. The record of its making does. A generative model can produce a competent poster, a competent rationale, and a competent reflection. It cannot produce the photograph of the twenty thumbnails a student drew, the critique they gave a named classmate on a Tuesday, or the revision they made in response to it.

This audit reads the rubric criteria of every graded item, not the assignment titles, and reports which of six kinds of process evidence each one already collects.

2 items collect little or none, which means the grade rests on an artifact alone.

strongPhotographic Postcard, Step 3: Refine15 pts
Physical captureProcess artifactsCritique givenRevision from named critiqueArticulated rationaleFile-level evidence
• Bind the prompt to something situated: the student’s own reference photographs, their own chosen subject, the specific critique they received.
strongSocial Media Animation, Step 1: Sketch15 pts
Physical captureProcess artifactsCritique givenArticulated rationaleSituated content
• Add a criterion for revision in response to a named critique. The critique is already happening; grading the response to it converts an exchange into evidence.
strongSocial Media Animation, Step 3: Refine15 pts
Process artifactsCritique givenRevision from named critiqueArticulated rationaleFile-level evidence
• Bind the prompt to something situated: the student’s own reference photographs, their own chosen subject, the specific critique they received.
strongGraphic Design Poster, Step 1: Sketch15 pts
Physical captureProcess artifactsCritique givenArticulated rationaleSituated content
• Add a criterion for revision in response to a named critique. The critique is already happening; grading the response to it converts an exchange into evidence.
strongPhotographic Postcard, Step 1: Sketch15 pts
Physical captureProcess artifactsCritique givenArticulated rationaleSituated content
• Add a criterion for revision in response to a named critique. The critique is already happening; grading the response to it converts an exchange into evidence.
strongPhotographic Postcard, Step 2: Build15 pts
Physical captureProcess artifactsCritique givenSituated contentFile-level evidence
• Add a criterion for revision in response to a named critique. The critique is already happening; grading the response to it converts an exchange into evidence.
• Add a criterion asking the student to name what changed and why, in the vocabulary of the discipline. One sentence is enough, and it is the cheapest AI-resistant criterion available.
• Admit a second submission mode. A recorded walkthrough of the decisions made is the single strongest addition here, and it is graded against the same criteria.
strongGraphic Design Poster, Step 3: Refine15 pts
Process artifactsCritique givenRevision from named critiqueArticulated rationaleFile-level evidence
• Bind the prompt to something situated: the student’s own reference photographs, their own chosen subject, the specific critique they received.
• Admit a second submission mode. A recorded walkthrough of the decisions made is the single strongest addition here, and it is graded against the same criteria.
strongSocial Media Animation, Step 2: Build15 pts
Process artifactsCritique givenSituated content
• Add a criterion for revision in response to a named critique. The critique is already happening; grading the response to it converts an exchange into evidence.
• Add a criterion asking the student to name what changed and why, in the vocabulary of the discipline. One sentence is enough, and it is the cheapest AI-resistant criterion available.
strongGraphic Design Poster, Step 2: Build15 pts
Process artifactsCritique givenSituated contentFile-level evidence
• Add a criterion for revision in response to a named critique. The critique is already happening; grading the response to it converts an exchange into evidence.
• Add a criterion asking the student to name what changed and why, in the vocabulary of the discipline. One sentence is enough, and it is the cheapest AI-resistant criterion available.
• Admit a second submission mode. A recorded walkthrough of the decisions made is the single strongest addition here, and it is graded against the same criteria.
adequateCourse Reflection10 pts
Physical captureProcess artifactsCritique given
• Add a criterion for revision in response to a named critique. The critique is already happening; grading the response to it converts an exchange into evidence.
• Add a criterion asking the student to name what changed and why, in the vocabulary of the discipline. One sentence is enough, and it is the cheapest AI-resistant criterion available.
• Bind the prompt to something situated: the student’s own reference photographs, their own chosen subject, the specific critique they received.
formativeIndustry Inspiration: Photography · 5 pts. Low-stakes completion work. Its function is participation, not measurement, and a model completing it costs the course nothing.
adequateTechnology Orientation10 pts
Critique given
• Add a criterion for revision in response to a named critique. The critique is already happening; grading the response to it converts an exchange into evidence.
• Add a criterion asking the student to name what changed and why, in the vocabulary of the discipline. One sentence is enough, and it is the cheapest AI-resistant criterion available.
• Require a process artifact: thumbnails, a storyboard, drafts, or a working board. The artifact alone is the part a model can supply.
adequateBehance Portfolio10 pts
Situated content
• Add a criterion asking the student to name what changed and why, in the vocabulary of the discipline. One sentence is enough, and it is the cheapest AI-resistant criterion available.
• Require a process artifact: thumbnails, a storyboard, drafts, or a working board. The artifact alone is the part a model can supply.
formativeIndustry Inspiration: Animation · 5 pts. Low-stakes completion work. Its function is participation, not measurement, and a model completing it costs the course nothing.
adequateLesson and Exercise: After Effects10 pts
Situated contentFile-level evidence
• Add a criterion asking the student to name what changed and why, in the vocabulary of the discipline. One sentence is enough, and it is the cheapest AI-resistant criterion available.
• Require a process artifact: thumbnails, a storyboard, drafts, or a working board. The artifact alone is the part a model can supply.
• Admit a second submission mode. A recorded walkthrough of the decisions made is the single strongest addition here, and it is graded against the same criteria.
adequateLesson and Exercise: Photoshop10 pts
Situated contentFile-level evidence
• Add a criterion asking the student to name what changed and why, in the vocabulary of the discipline. One sentence is enough, and it is the cheapest AI-resistant criterion available.
• Require a process artifact: thumbnails, a storyboard, drafts, or a working board. The artifact alone is the part a model can supply.
• Admit a second submission mode. A recorded walkthrough of the decisions made is the single strongest addition here, and it is graded against the same criteria.
formativeStudent Success: Reputable Tutorial/Learning Sources · 3 pts. Low-stakes completion work. Its function is participation, not measurement, and a model completing it costs the course nothing.
exposedLesson and Exercise: Illustrator10 pts
File-level evidence
• Add a criterion asking the student to name what changed and why, in the vocabulary of the discipline. One sentence is enough, and it is the cheapest AI-resistant criterion available.
• Require a process artifact: thumbnails, a storyboard, drafts, or a working board. The artifact alone is the part a model can supply.
• Bind the prompt to something situated: the student’s own reference photographs, their own chosen subject, the specific critique they received.
exposedColor Palette Exercise10 pts
Situated content
• Add a criterion asking the student to name what changed and why, in the vocabulary of the discipline. One sentence is enough, and it is the cheapest AI-resistant criterion available.
• Require a process artifact: thumbnails, a storyboard, drafts, or a working board. The artifact alone is the part a model can supply.
• Admit a second submission mode. A recorded walkthrough of the decisions made is the single strongest addition here, and it is graded against the same criteria.
formativeStudent Success: Campus Cares · 3 pts. Low-stakes completion work. Its function is participation, not measurement, and a model completing it costs the course nothing.
formativeStudent Success: Digital Media Programs at the college · 3 pts. Low-stakes completion work. Its function is participation, not measurement, and a model completing it costs the course nothing.
formativeIntroduce yourself · 5 pts. Low-stakes completion work. Its function is participation, not measurement, and a model completing it costs the course nothing.
formativeIndustry Inspiration: Graphic Design · 5 pts. Low-stakes completion work. Its function is participation, not measurement, and a model completing it costs the course nothing.
formativeStudent Success: Advising · 3 pts. Low-stakes completion work. Its function is participation, not measurement, and a model completing it costs the course nothing.
formativeSyllabus and Course Introduction · 5 pts. Low-stakes completion work. Its function is participation, not measurement, and a model completing it costs the course nothing.
formativeStudent Success: Where to get help · 3 pts. Low-stakes completion work. Its function is participation, not measurement, and a model completing it costs the course nothing.
formativeStudent Success: Online Learning Tips · 3 pts. Low-stakes completion work. Its function is participation, not measurement, and a model completing it costs the course nothing.

What to do
  1. Grade the process you already require. Most courses ask for sketches, drafts, and critique and then grade only the final file. The evidence is being produced and discarded. A criterion is all that is needed to capture it.
  2. Revision in response to a named critique is the strongest single criterion available, and it costs the student nothing new. They are already receiving feedback. Grade what they did with it.
  3. Authentic does not mean AI-resistant. A model will write a portfolio case study, a client rationale, and a reflective essay. Authenticity of task is not the defence; process evidence is.
  4. A recorded walkthrough converts an artifact into a performance. Two minutes of the student explaining their own decisions is un-fakeable and is graded against criteria that already exist.
Seat hours and workloadCredit-hour policy
review

Estimated student workload · 30.8 hours, which is 116% of the 26.5-hour outside-work allowance

Course-wide

1 credits × 40 hours = 40 total student hours. That figure is fixed by the credit, not by the load. Load hours change how the total is split between direct instruction and outside work; they never change the total.

Reading, 16,105 words across 34 pages, at 200 wpm1.3 h
Video, 8 videos at 6 minutes assumed0.8 h
Graded work, 27 items scaled by point weight28.6 h
Total30.8 h

Estimate. This figure is only as good as its assumptions. Video length is assumed because the export stores links, not durations. Effort per graded item is inferred from point weight, which is a proxy, not a measurement. Treat the figure as a starting position for a curriculum conversation, not a finding.

Can a student use it?

Accessibility, usability, and learner variability.

Accessibility. It verifies the structural properties recoverable from the package: alternative text, descriptive link text, table headers, heading order, typeface, and captions required on video. Contrast is verified against WCAG AA at the moment a palette is chosen, so a failing pair cannot be applied at all.

Limits. It does not open PDFs, and it will not write your alternative text. Alt text depends on why a given image is on a given page, which is not recoverable from a file, and a generated caption describes the picture instead of its function while looking finished. The instrument locates the images and leaves the sentence to the instructor.

For PDFs, use YuJa Panorama. The district licenses it and it is enabled in Canvas. It appears as a gauge beside every file, performs text recognition on scans, tags headings and reading order, and generates alternative formats a student can download without disclosing a disability to anyone. A scanned PDF is an image of text: a screen reader announces nothing, and the failure is invisible to a sighted instructor because the page looks correct.
Universal Design for LearningCAST framework
pass

Universal Design for Learning is not accessibility

How this section works

Accessibility (WCAG) removes barriers for learners with disabilities. It is a compliance obligation.

Universal Design for Learning designs for learner variability from the outset. It is a pedagogical framework.

A course can be fully WCAG-conformant and still offer a single narrow path through it. Conflating the two is the most common error in the field.

review

Principle 3 · 19 of 27 graded items accept a single submission format

Course-wide

Multiple Means of Action and Expression. Canvas records the submission modes each assessment accepts. Here, 70% of graded items permit exactly one.

A learner who could demonstrate the same competency through a recorded walkthrough, an annotated artifact, or an oral defense has no route to do so. Where submission format is not itself the construct under assessment, it is acting as a barrier.

What to do
  1. For each single-mode item, ask: is the format the competency, or merely the container? A typography assignment must be submitted as a file. A rationale need not be typed.
  2. Where it is the container, admit a second mode. Canvas: open the assignment → Submission TypeOnline → check more than one of Text Entry, Website URL, Media Recording, File Upload.
  3. Rubric first. A rubric that says “well-written paragraph” cannot score a spoken defense. Rewrite criteria to describe the competency, not the artifact, and then both modes become gradeable against the same standard.
  4. Start with the reflective and explanatory items. They are where the format is least load-bearing and the barrier is highest.
review

Principle 2 · content is offered in 2 modalities

Course-wide

Multiple Means of Representation. Computed from the package: 34 text pages, 69 video references, 2 images, 0 audio references, 0 print/PDF resources.

There is no audio-only pathway. A commuting learner has no way to engage this course without a screen.

There is no printable alternative.

What to do
  1. Neither gap is a compliance failure. Both are options a learner does not currently have, and adding one is cheap.
  2. Audio: if your videos already have accurate captions, the transcript is written. Reading it aloud, or exporting the audio track alone, gives you an audio pathway at near-zero cost.
  3. Print: Canvas pages export to PDF from the browser print dialog. One PDF per module, linked in the module overview, covers it.
  4. Do not add modalities for their own sake. The coherence principle (Mayer) holds that extraneous material depresses learning. Add a modality when it opens a path, not when it decorates one.
pass

Principle 1 · 51 expressions of learner choice

Course-wide

Multiple Means of Engagement. Counted across all course text.

Choice language is substantive rather than token. Where learners select their own subject and carry it across the course, this is genuine autonomy, and it is also what makes the work resistant to generative substitution: a model cannot have been to the place the student chose.

What to do
  1. Give learners one durable choice they carry across the whole course rather than many trivial ones. A subject, a client, a destination, a dataset.
  2. The durable choice does the double duty: it drives engagement (self-relevance) and it makes every downstream artifact un-generatable, because the model has no access to the thing they chose.
  3. Distinguish meaningful choice from nominal choice. Picking one of three instructor-supplied prompts is not autonomy. Choosing the thing the work is about is.
pass

What this check cannot evidence

Requires human review

Three UDL indicators are recoverable from a package: accepted submission modes, content modality, and the presence of choice. The remainder are not. Whether options are meaningful rather than nominal, whether scaffolds support self-regulation, and whether the course sustains effort across a term are matters of professional judgment. They are surfaced for review, never scored.

Course overview and informationOSCQR 1-10
pass

OSCQR 1-10 · all seven elements of course information were found

Course pages

a Start Here or course overview page; instructor contact information; stated response and turnaround times; a statement of required technology; technical support routes; accessibility and accommodation routes; expectations for participation.

Detected by keyword, so confirm the substance rather than the presence.

Technology and supportOSCQR 11-15
fix

OSCQR 14 · no privacy policy linked for 5 third-party tools

Course-wide

Third-party tools referenced by this course:

• Discord, 35 mentions, no policy linked
• Miro, 55 mentions, no policy linked
• Behance, 28 mentions, no policy linked
• Adobe, 78 mentions, no policy linked
• YouTube, 65 mentions, no policy linked

Students are required to create accounts on commercial platforms as a condition of passing this course. They are entitled to know what those companies do with their data, and a reviewer will flag its absence immediately.

What to do
  1. Create one page in the first module. Call it “The Tools We Use, and What They Know About You.”
  2. Link each policy directly:
      • Discord
      • Miro
      • Behance
      • Adobe
      • YouTube
  3. For each, write one sentence on what the student is agreeing to. Not legalese.
  4. State your fallback. A student who declines an account on a commercial platform still needs a route through your course. Name it. A reviewer will ask.
Design and layoutOSCQR 16-28
pass

OSCQR 23 · 871 font declarations, all sans-serif

Course-wide

“A sans-serif font with a standard size of at least 12 pt is used.”

All declared typefaces are sans-serif, satisfying the standard.

pass

OSCQR 36 · 2 images, all carrying alternative text

Course pages

Every image carries an alt attribute, including decorative images marked alt="".

pass

OSCQR 21 · heading structure is sequential on all 34 pages that use headings

Course pages

Headings descend one level at a time on every page that uses them, so the document outline is navigable.

pass

OSCQR 37 · all link text is descriptive

Course-wide

No “click here” links. This is rarer than it should be.

Content and activitiesOSCQR 29-37
review

OSCQR 35 · 8 videos to caption · 2 embedded, 6 linked (25 channel links, not videos)

Course-wide

2 embedded and 6 linked videos carry course content and require captions. A further 25 YouTube channel links were found and are not counted: a channel is a resource list, not a video, and there is nothing on it to caption.

Embedded:
• Canvas Studio · d0e032e5
• Canvas Studio · bd2fae4f

Linked:
https://www.youtube.com/watch?v=7go-QJ5Z64A
https://www.youtube.com/watch?v=YeI6Wqn4I78
https://youtu.be/3BSswUDrIjg
https://youtu.be/A8eM7mdk0_k
https://www.youtube.com/watch?v=eaOoqqmckXw
https://www.youtube.com/watch?v=QrNi9FmdlxY

The export stores video references, not video files, so no tool can verify captions from here. Open each one.

The requirement is WCAG 2.1 SC 1.2.2, which mandates captions, not transcripts. Quality is specified by the DCMP Captioning Key. Auto-captions satisfy neither: they run 70 to 90 percent accurate and degrade worst on discipline-specific terminology, which is exactly the content being taught.

What to do
Editing auto-captions IS the standard workflow, and it does satisfy the requirement. You do not need a captioning vendor. You need twenty minutes per video.
First, determine which case you are in. The two cases share no steps, and most faculty guess wrong.
  1. Open the video. Does it play from Canvas Studio, or is it on a YouTube channel you control? Then it is yours, and you can caption it.
  2. Is it embedded or linked from Adobe, a studio, or another creator? Then it is not yours, and you cannot caption it and may not re-upload it.
Case A · Your video, in Canvas Studio
  1. Canvas → Studio in the global (left) navigation → open the video.
  2. Captions tab → select EnglishRequest Captions. Studio generates them in a few minutes.
  3. Now the step that matters, and the step nearly everyone skips. Click Edit and read the track against your audio. Fix the technical vocabulary first: it is what auto-captioning destroys and it is what you are teaching. Keyframe becomes key frame. Bézier becomes anything at all. Bleed, kerning, comp, opacity, alpha, check every one.
  4. Publish. The captions now travel with the video into every course copy.

Yes, this counts. WCAG does not require a human to type from scratch. It requires captions that are accurate and complete. An auto-caption you have corrected is a human caption.

Case A2 · Your video, on YouTube
  1. studio.youtube.comContent → click the video.
  2. Left sidebar: Subtitles. There will already be a track labelled English (automatic).
  3. Click Duplicate and Edit beside it. This is the whole trick, it converts the auto track into an editable one instead of making you start over.
  4. Play through, correct the terminology, and Publish.
  5. The track is now a real caption track. The “automatic” label is gone, and that label is exactly what a reviewer looks for.
Case B · Someone else’s video. This is where people get stuck.

You cannot add captions to a video you do not own. And you may not download it and re-upload it to caption it, that is a copyright violation, and it is the remedy faculty most often reach for without realising. Your options, best first:

  1. Check whether it already complies. Play it and turn on CC. If the track says English (auto-generated), it does not satisfy the standard. If it says simply English, the owner captioned it and you are done. Many Adobe and major-channel tutorials already have real captions. Check before you do anything else, this resolves it more often than people expect.
  2. Find a captioned equivalent. On YouTube, search your topic and apply the Subtitles/CC filter under Filters. Faster than everything below it.
  3. Use the vendor’s own documentation. helpx.adobe.com is professionally captioned and it will not be deleted out from under your course in two years, which a random YouTube tutorial absolutely will be. For a software demonstration this is usually the correct answer.
  4. Ask the owner. A comment or an email asking them to add captions costs five minutes and sometimes works.
  5. Re-record it yourself. Not the burden it sounds like. A three-minute screen capture in Studio, in your own voice and your own course vocabulary, is usually better pedagogy than a stranger’s video and it resolves the problem permanently.
  6. Amara (amara.org) can caption third-party video without re-uploading it. Note its free public workspace closed in April 2026; it now runs roughly $5 per video.
What does NOT satisfy this standard
  1. A transcript is not a caption. WCAG 1.2.2 requires captions specifically. A deaf student reading a transcript in a separate window cannot watch your cursor and read the words at the same time, which is the entire point of a software demonstration.
  2. Auto-captions left uncorrected are not captions. They are a first draft. 70 to 90 percent accurate, and they fail hardest on the exact words the video exists to teach.
  3. “It has CC available” is not the same as “it is captioned.” Look at the track label.
pass

OSCQR 34 · no PDFs; all text content is HTML

Course Files

The course contains zero PDFs. Every text resource is HTML, which reflows, scales, and is natively readable by assistive technology. This is the correct answer and most courses do not achieve it.

InteractionOSCQR 38-44
pass

OSCQR 39-42 · peer interaction runs outside Canvas

Course-wide

The course contains no Canvas discussions, and its pages reference an external venue (Discord, Slack, or a studio critique). That is a legitimate design and this check does not penalise it.

The consequence worth knowing: Canvas holds no record of the interaction. If asked to evidence Regular and Substantive Interaction, the evidence lives in the other platform and is the instructor’s to produce.

Assessment and feedbackOSCQR 45-50
review

OSCQR 46 · 10 of 27 graded items carry no rubric

Assignments

• Student Success: Reputable Tutorial/Learning Sources (3 pts)
• Industry Inspiration: Graphic Design (5 pts)
• Industry Inspiration: Photography (5 pts)
• Student Success: Advising (3 pts)
• Student Success: Campus Cares (3 pts)
• Syllabus and Course Introduction (5 pts)
• Student Success: Where to get help (3 pts)
• Student Success: Online Learning Tips (3 pts)
• Industry Inspiration: Animation (5 pts)
• Student Success: Digital Media Programs at the college (3 pts)

A grade without criteria is a number the student cannot act on. A rubric is also the only structure Canvas offers for binding an outcome to an assessment, so an item with no rubric can never evidence a competency or an ILO.

What to do
  1. Canvas: open the assignment → + Rubric. Reuse an existing one with Find a Rubric rather than authoring each from scratch.
  2. Write criteria that name the performance, not the effort. “Composition reads at three metres” is a criterion; “shows effort” is not.
  3. Once the rubric exists, bind the outcome: Find Outcome on the criterion that assesses it.
pass

OSCQR 45 · 27 graded items, 248 points, 18 worth 10 points or fewer

Assignments

The course distributes assessment across 27 items rather than concentrating it. 18 are low-stakes, which gives a student a way to find out where they stand before the cost of being wrong is high.

Fix, and prepare to reimport

Institutional outcomes, visual system, dates, and course Files.

Institutional Learning Outcomes
pass

Institutional Learning Outcomes · purpose and scoring standard

How this section works

Institutional Learning Outcomes are how your college proves its students can do the things a degree claims they can do. Every student’s attainment is tracked across their whole time in college, first semester to last.

They are not course points. An ILO score rides alongside the grade and is never used to evaluate an individual student or a faculty member. A student can earn an A and score Emerging, and nothing is wrong.

And here is the rule that is almost always applied backwards. The ILO reads “by the end of their college experience…”, so you score against a graduating student, not the one in front of you.

Work that is proficient for your course may be emerging for the ILO. A 100-level student should mostly score Emerging (2). This is the expected distribution, not a deficiency. It is how the institution sees growth over time. If every intro course reports Proficient, the data is worthless and no one can prove students learned anything.

Score honestly. Score low. That is the assignment.

pass

One outcome, one criterion, one assignment

How this section works

Canvas calculates ILO mastery as Highest Score. So attaching an outcome to nine rubrics does not make it more rigorous, it gives more chances at the same number, and it burdens the rubric with criteria that will not be scored.

Attach each ILO once, to the single criterion where the evidence is strongest, late in the course, where the student is at their best.

Three to six ILOs is a well-instrumented course. More than that is theatre.

review

6 of 47 Institutional Learning Outcomes are evidenced by work this course already grades

Canvas → each assignment → Rubric → Find Outcome

Visual Communication 3
VC1VC2VC3VC4VC5
Oral Communication
OC1OC2OC3OC4OC5
Written Communication
WC1WC2WC3WC4WC5
Critical Thinking
TI1TI2TI3TI4TI5
Creative Thinking 2
TE1TE2TE3TE4TE5TE6
Information Literacy
I1I2I3I4I5I6I7I8I9
Career Goals & Workplace 1
W1W2W3W4W5W6W7
Personal & Community Well-Being
WB1WB2WB3WB4WB5WB6
Diversity
D1D2D3D4
Quantitative Reasoning
Q1Q2Q3Q4Q5Q6

6 of 58 institutional outcomes are evidenced by work this course already grades. evidenced here. assessed elsewhere in the college.


Each highlighted outcome is work you are already grading. The rubric simply never said what it was for. Attaching one changes nothing about how you teach, and it puts your learners into the institution’s record of what they can do.
The six, and where each one attaches
ILO-VC5 · Analysis: Critique & Interpretation
Find Outcome → Communication Visual
“Analyzes, critiques, and interprets work from a visual, historical, or conceptual framework.”
Attach to: The critique criterion on your last graded project of the term.
The highest-value attachment in most studio courses. Critique frequently constitutes 20 percent of a rubric and maps to no course competency. VC5 names it exactly. Use the last critique of the term, not the first, you want their most mature analysis.
ILO-TE1 · Acquiring: Evaluates Creative Process
Find Outcome → Creative Thinking
“Evaluates creative process and product using domain-appropriate criteria.”
Attach to: The criterion where a learner articulates what they changed after critique, and why.
If your rubric asks a student to name the principle behind a revision, you are already assessing this and not claiming it.
ILO-TE2 · Risks: Seeks Untested Approaches
Find Outcome → Creative Thinking
“Actively seeks out and follows through on untested and potentially risky directions… may include personal risk or risk of failure.”
Attach to: Any project criterion. Consider adding one.
Read this one twice. Most studio rubrics punish creative risk: a learner who attempts something ambitious and lands it at 80 percent scores below one who plays it safe and lands at 95. TE2 is institutional permission to reward the ambitious attempt that does not fully land, and it costs nothing to attach.
ILO-W4 · Job Search: Marketing Self to Employers
Find Outcome → Career Goals & Workplace Skills
“Ability to market themselves and clearly communicate how classes… translate into qualifications for possible careers.”
Attach to: The published portfolio, and specifically its written description.
That criterion describes a portfolio without ever using the word. If your course ends in a published body of work, this attachment is free.
ILO-VC4 · Delivery: Visual Production
Find Outcome → Communication Visual
“Makes compelling choices regarding the delivery of a final visual production; choices reflect a physical or digital environment.”
Attach to: The criterion where a file is prepared for output, resolution, color mode, bleed, delivery size.
Print versus screen is “physical or digital environment.” If you grade whether the file would actually print, you are assessing this.
ILO-VC3 · Execution: Technical Skills
Find Outcome → Communication Visual
“Technically executes an idea… selects appropriate equipment, software, peripherals, and workflow.”
Attach to: The build step with the fullest toolchain, or the criterion assessing how physical work was captured.
“Equipment and peripherals” covers the camera or scanner a learner uses to bring paper work into a digital workflow. Most courses do this and grade only the artifact, never the capture.
Why this matters, and how to score it

An institutional outcome is the college’s answer to the question a degree implies: can our graduates communicate visually, think creatively, use information responsibly, find the support they need? Attainment is tracked across a student’s whole time in college, and a course contributes to that record only where an outcome is bound to a rubric criterion on graded work.

An outcome assessed only inside the major produces data describing majors alone. An introductory course full of non-majors is the most consequential place to attach one, because for many of those students it is not their first data point. It is their only one.

Score against a graduating student, not the one in front of you. Work that is proficient for your course may be emerging against a terminal standard, and emerging is the correct rating at the introductory level. That is the signal, not a deficiency: it is how growth across a degree becomes visible at all. Uniform proficiency in introductory courses renders the longitudinal data uninformative, and the institution loses its ability to demonstrate that its students grew.

Canvas calculates mastery as Highest Score, so one attachment per outcome is sufficient. ILO scores are not grades. They do not affect a student’s standing and are never used to evaluate a student or an instructor.

What to do
How to attach an ILO to a rubric. About two minutes, once.
  1. Open the assignment you want to attach it to.
  2. Scroll to its Rubric and click the pencil to edit. If it has no rubric, create one first. An ILO attaches to a rubric criterion, not to an assignment.
  3. Click Find Outcome.
  4. In the left column, open the domain folder named above.
  5. Select the outcome, read its description to be sure, and click Import.
  6. It is now a criterion on your rubric, scored 0 to 4. Save.
Do not see Outcomes in your course navigation?
  1. SettingsNavigation tab.
  2. Drag Outcomes into the upper list. Save.
Consult your institutional assessment office
  1. Where to ask: ILO Faculty Guide, see “Importing ILO rubric items into Canvas.”
Visual consistency
review

The course uses 2 typefaces, 8 font sizes, and 5 colours

Course-wide

34 of 34 pages carry inline styling.

Typefaces: verdana, geneva, sans-serif, verdana,geneva,sans-serif
Sizes: 12pt, 14px, 15px, 16px, 17px, 19px, 22px, 24px
Colours: #ffffff   #22332c   #2f6b52   #7a5a1f   #e3efe9

A page that changes typeface or scale between sections asks the student to re-learn the interface each time. Consistency is not decoration; it is a reduction in extraneous cognitive load, which is the load that teaches nothing.

What to do
  1. Choose one style below and apply it. The tool rewrites every page inline, because the Canvas sanitizer strips <style> blocks and class-based CSS. Inline is the only styling that survives the editor.
  2. Review the preview before exporting. The preview renders exactly what the package will contain.
Due dates and deadline load
review

1 date carry more than two deadlines

Course calendar

2026-08-27 · 3 items
Syllabus and Course Introduction
Student Success: Digital Media Programs at the college
Introduce yourself


Deadline stacking is the most common cause of a late-work spike. Students triage, and the item they drop is the one worth fewest points, which is often the one that scaffolds the next unit.

What to do
  1. Move the lowest-stakes item on each stacked date. Canvas: Syllabus shows every date in one table; edit from there rather than opening each assignment.
  2. Aim for no more than two deadlines on a day, and keep the pattern predictable across weeks so students can plan.
Course Files
pass

No student data found in course Files

Course Files

A course export excludes enrollments, grades, submissions, and discussion replies. But it includes everything ever uploaded to course Files, and that is where student data ends up.

Nothing in this course’s Files looks like a roster, a gradebook, or a class list.

OSCQR, the SUNY Online Course Quality Review Rubric, is made available under a CC BY 4.0 license by the State University of New York. This instrument applies the rubric and is not affiliated with or endorsed by SUNY.