25%, Individual or pair. Due Week 12. Screened Week 13.

Digital Storytelling Project

You make a 3-to-5-minute video on one of three tracks, and every track ends with something a person outside this class could watch.

Assignment

Pick a track before anything else, because the three are graded on the same rubric but prepared in different ways. Track A is your own story, the path that brought you into this work. Track B explains a social problem to an audience that could act on it. Track C is a piece an organization could use, about one of its programs or its mission. The next section states what each one requires; everything after that section applies to all three unless it says otherwise.

Track B is the course’s default and the one most of the page is written for. This is not an awareness video. A meta-analysis of 39 entertainment-education studies finds a standardized effect of 0.48 on knowledge, 0.18 on attitudes, 0.17 on intentions, and 0.16 on behavior, so the effect shrinks as the outcome becomes more consequential. The rubric grades the audience you name, the action you ask that audience to take, and the reason you give them to think the action will work.

Work solo or in a pair. Pairs make one piece and share one reflection, with a line on who did what, and a pair needs a problem both people can claim some standing on. Pick a partner whose production skills complement yours; the pair option exists for scope, since two people can handle filming and editing that would swamp one.

Download the working template, with the panel table and the 5-field reflection: SIL503_storyboard.docx

Production has two lanes: an easy cloud lane for most footage, and an offline lane required once your footage includes an identifiable vulnerable person. The Tools page describes both, with the specific free software for each.

Cameras, mics, and lights come free from North Campus Technology Checkout, and the School’s own podcast studio and the Shapiro Design Lab video room are both bookable by students; the School of Social Work resources in the syllabus lists all of them with the booking steps. Book early, because Week 11 is when every media course on campus wants the same room.

Three tracks

Track A, a personal story

A first-person piece about your own path into this work: what you came from, what you decided, and what you are trying to do now. StoryCenter defines the form as a two to four-minute first-person video built from a narration script of about 300 words, with recorded voice over still and moving images, and its public gallery is where to watch finished ones before you write yours.

The structure is Marshall Ganz’s public narrative, which Week 10 covers as a method: a Story of Self is built from a challenge, a choice, and an outcome, and it is told to convey a value rather than to report a biography. Going on to the Story of Us and the Story of Now is optional here, and the sheet below asks you to say where they would go.

This track drops the named audience segment, the behavioral ask, and the efficacy claim, and replaces the communication brief with the story of self sheet. That shortens the planning without lowering the standard, because it moves the weight onto the 20-point ethics row. A story about your own path almost always contains family, clients, colleagues, and supervisors who did not choose to be in it, the consent rules below apply to every one of them, and the reflection asks whom you omitted and why.

Track B, explainer

The problem is explained to people positioned to act on it. This is the course default and the rest of this page is written for it: the problem comes from your own Problem Inventory entries, something you watched happen at a job, a volunteer role, a board you serve on, or from your own experience of a system, alone or with family, and your standing to make the piece rests on that entry. The ABT requirement, the seven-field communication brief, a named audience, and a specific behavioral ask all apply.

Track C, organizational piece

A piece a real organization could use, about one of its programs or its mission, made to a brief that organization would recognize. Footage comes from one of two places: a community organization you already have a relationship with, or the Edward Ginsberg Center’s footage archive, which is available to this course by request through me. Ask in Week 9 at the latest, because the request is not self-serve and the archive has to be searched for material relevant to your subject.

This track is built on a client relationship, and that is also where the difficulty lies. The organization owns the message, you own the edit, and the people in the footage chose neither. Consent and representation take most of the weight here, and the readings on the single story and on inspiration porn are required before you script rather than after. The communication brief applies with one change: the audience and the ask come from the organization rather than from you, and field 7 asks what the organization would lose if the piece went wrong.

Evidence behind the ask

Tracks B and C. Track A asks a viewer to understand something rather than to do something, so the campaign evidence below governs the other two.

Wakefield, Loken, and Hornik’s review of mass media health campaigns concludes that concurrent access to services and products is what converts a motivated viewer into an acting one, and that campaigns routinely “address behaviours that audiences lack the resources to change.” Check that the action you are about to ask for exists and that your audience can reach it.

Webb and Sheeran found that interventions producing a medium-to-large shift in intention produced only a small-to-medium shift in behavior. Gollwitzer and Sheeran’s meta-analysis of 94 tests found that an implementation intention, an if-then plan naming when, where, and how, produced a medium-to-large effect on goal attainment over and above intention alone. This is the reason the brief asks for a time and a place.

Tannenbaum and colleagues’ synthesis of 127 papers and 248 samples found fear appeals effective on average, and more effective when the message included efficacy statements. A piece that shows how bad something is still owes the viewer a reason to believe their action will matter.

Västfjäll, Slovic, and Mayorga documented pseudoinefficacy: making salient the people who cannot be helped reduces giving toward the person who can be, and Slovic’s earlier work on psychic numbing finds that the collapse of compassion can begin at two. Opening on the size of a problem can therefore work against the ask that closes the piece.

Christiano and Neimand’s “Stop Raising Awareness Already” is the practitioner version of all of this and a required reading. It names four risks of awareness-only communication and prescribes narrow targeting, a specific behavioral ask, an explicit theory of change, and trusted messengers, and its companion piece, How to Build Better Calls to Action, examines what makes an ask specific, achievable, easy, and normative.

Cho and Salmon catalogue eleven types of unintended effect in health and social communication, including boomerang, desensitization, culpability, social norming, and opportunity cost. The reflection asks which of the eleven your piece is most exposed to, so read the list while you are still scripting.

ABT check

Tracks B and C. Track A’s equivalent is the challenge, choice, and outcome line on the story of self sheet.

Before you write a script, you write one sentence in Randy Olson’s ABT form: ”___ and ___, but ___, therefore ___.” And is setup, but is the conflict, therefore is the consequence.

The sentence has to be approved before you start scripting, and it also works as a scope check, since a topic where you cannot complete the “but” has no conflict in it and is too broad for a 3-minute piece. “Homelessness is bad and we should care about it” supplies nothing to follow the but. “Our county funds rapid rehousing and the waitlist software assumes a working phone number, but a third of applicants change numbers within sixty days, therefore the people highest on the list are the ones the system loses” yields a video.

Communication brief

Tracks B and C. The Week 10 deliverable is a one-page brief with seven fields; Track A submits the story of self sheet instead, on the same date.

1. The problem, in the four-field register of the Problem Inventory: the situation, who is affected, why it persists, whether technology is relevant.
2. The audience, named as narrowly as you can stand, with a reason you believe this particular group can act. “The public” is not an audience. Maibach and colleagues’ Six Americas study shows how much sharper a real segmentation is than a demographic.
3. The change you want in that audience, stated as a before and after.
4. The behavioral ask, observable, and specific about when, where, and how. “Call your county commissioner’s office during the public comment window on November 4, using the three sentences in the description” offers the viewer a specific action on a particular day. “Learn more” does not.
5. The efficacy claim. Why should this audience believe the ask will work? What evidence do you have that it has worked before?
6. The distribution channel. Where is the piece seen, and by that segment specifically? Craft cannot compensate for a video the intended audience never encounters.
7. The assumption most likely to be false. Name the weakest link in the chain from viewing to outcome. This field is graded on honesty, and “my audience already agrees with me and will do nothing anyway” is a legitimate answer that should send you back to field 2.

Fields 3 through 5 are a theory of change compressed to a paragraph. If you want the long form, the Community Tool Box on writing objectives and Spitfire’s Smart Chart 3.0 are both free and built for exactly this.

Sample brief

Problem: at a food pantry I volunteered with, clients who missed two appointments were auto-closed out of the program; most misses came from reminder emails going to addresses people had lost access to. Audience: the twelve to twenty staff and board members of small pantries in Washtenaw County who set their own no-show policy, reachable through the county food-security coalition’s monthly call. Change: from treating the no-show policy as a fixed administrative fact to treating it as a decision they own. Ask: put “review the no-show and case-closure policy” on the agenda of your next staff meeting, using the one-page discussion guide linked in the description. Efficacy claim: two pantries in the coalition changed the policy after a member raised it; both report serving more people, and one of their directors says so on camera. Channel: screened at the coalition’s monthly call, with the link circulated on its listserv. Weakest assumption: that the policy is the pantry’s and not with the funder’s reporting requirements. If it proves to be the funder’s, the audience in field 2 is wrong.

Every later decision on the project is measured against this page, so it is worth the hour it takes.

Story of self sheet

Track A submits this page in place of the communication brief, at the same deadline.

1. The value. The single idea you want a viewer to understand about what you care about and why. A value, not a topic.
2. The challenge. A specific moment when something was at stake for you. A moment has a date, a place, and other people in it; a period of your life does not.
3. The choice. The decision you made, and the live alternative you had at the time. A choice with no alternative behind it is a description.
4. The outcome. What followed, including the case where nothing resolved.
5. Who else is in it. Everyone identifiable in the story, what they would say about appearing in it, and what you are asking of them.
6. Where Us and Now would go. One sentence each, even if the finished piece ends at Self.
7. The omissions. Everything you decided to omit from your own story, and the reason for each. Graded on honesty, like field 7 of the brief.

Standing and positionality

Almost the entire research literature on digital storytelling describes a participatory practice: people telling their own stories, with facilitation. Track A is that case, and its ethical work is about the other people inside your own story. Tracks B and C are the other configuration, where you are making a piece about a problem that affects other people and you control the edit. The ethical critique of issue media is aimed at that configuration, and working without a partner organization sharpens the problem instead of dissolving it.

The Problem Inventory requirement exists so that you can say you are describing a system you have worked inside, which is a position you can defend under questioning in a way that having read about a group is not. You also write a positionality paragraph, and it is graded. Gubrium, Hill, and Flicker name six recurring ethical tensions in this kind of work, and the one students consistently underestimate is the power of shaping: the maker’s ability to bend a story toward a predetermined message. Your paragraph answers three questions. Whose experience is this piece about? What standing do you have to narrate it? What would be different if the person most affected were present while you edited?

If the honest answer to the second question is “not much,” the usual fix is to retain the topic and change what the video is about, moving it toward the part of the system you have actually worked inside and away from the people the system acts on.

Watch Chimamanda Ngozi Adichie on the danger of a single story and Stella Young on inspiration porn before you script.

Consent and ethical representation

This criterion is worth 20 points, and the requirements below are hard.

Name and apply one ethics standard. Pick one document from the list below, apply it to your piece, and say in the reflection which one you used and how, since the points are for the application.

• Bond’s Statement of Ethical Practice, one page, with the full guidelines behind it. It treats informed consent as an ongoing dialogue and not a signed form.
• The Dóchas Guide to Ethical Communications (2025) for principles, and the older Illustrative Guide (2014) for worked good and bad examples, which is more concrete than the document that replaced it.
• UNHCR’s Ethical Communications Guidelines (2024), which ships a consent form template, consent checklists, and an inclusive-language glossary.
• WITNESS on conducting safe, effective, and ethical interviews with survivors, organized before, during, and after, and written for filmed interviews specifically.
• The Documentary Accountability Working Group’s framework, whose six values can be used directly as a self-audit.

Some topics have a mandatory standard on top of your chosen one. If your piece touches suicide or self-harm, the Recommendations for Reporting on Suicide, developed jointly by SAVE, AFSP, AAS, CDC, NIMH, and AP Managing Editors, are not optional, and the WHO resource for media professionals is the international companion. For sexual violence, use NSVRC’s reporting guidance; for disability, the NCDJ Disability Language Style Guide; and for immigration, Race Forward’s journalist style guide. Any interview that touches trauma is covered by the Global Center for Journalism and Trauma’s interviewing tip sheets, which tell you not to ask “how do you feel,” a question survivors consistently name as the most distressing one they are asked.

Right of reply. Any real person who appears in or is identifiably described by your video watches the cut before it screens and holds a veto or an edit. You document what changed, with a name and a date on that step of your production schedule. Consent is staged: a yes to being interviewed is not a yes to being in the cut, and a yes to the cut is not a yes to publication.

No synthetic depictions of real people, and no generated imagery of the people your problem affects. The first rule is the old one: generative AI may not depict real, identifiable people or communities as if the footage were real, with no exceptions. The second is new this year and follows the evidence. Bianchi and colleagues found that text-to-image models amplify demographic stereotypes at scale, that the stereotypes appear whether or not the prompt mentions identity, and that they survive both user counter-prompting and vendor guardrails. Bloomberg’s analysis of more than 5,000 Stable Diffusion images found “inmate” returning five darker-skinned images for every lighter-skinned one in a system where under half of US prisoners are people of color. Mack and colleagues, working with 25 disabled participants, found models returning reductive archetypes: manual wheelchairs, sadness, loneliness, inspiration. And The Lancet Global Health has a name for where this is heading, poverty porn 2.0: AI-generated images that reproduce the visual grammar of poverty while avoiding the consent and the cost of photographing anyone real, then are scraped to train the next model. Generated backgrounds, textures, abstract motion, and diagrams remain permitted. Generated images of the people your problem affects are not permitted on any topic.

Label and log what you generate. Any AI-generated image or audio in the piece is labeled inside the piece, where the viewer sees it, and every generated asset is logged in the reflection with the exact prompt used.

Beyond the rules, the reflection asks you to name where your video gives voice and where it could extract, and what you did about that. The points go to an honest answer, including one that admits the piece was wrong about something.

Forms

The form follows from the brief and from what you can ethically shoot, and every one of these can support a 3-to-5-minute issue piece.

• First-person testimony, one consenting person, told in their own words.
• Explainer, a structural argument built from narration plus graphics, archival material, or animation.
• Data-driven, where charts and numbers are the visual engine.
• Observational, where the camera watches a process or a place instead of a person.
• Essayistic, where an object, a document, or a designed experiment makes the argument.

The last three are the no-filming lane. If your topic makes filming a person risky or unfair, build the piece from public-domain archival footage, data visualization, your own object photography, or animation. When you do need a real voice but not a real face, WITNESS’s guide to concealing identity in interviews covers when to protect someone and how: backlight to silhouette, shoot only the hands, shoot deliberately out of focus, blur in post. If you animate an interview instead, know that animation is not ethically neutral either and be ready to defend the choice in the reflection.

Examples

Real issue videos, grouped by form. Runtimes and caption status were checked in August 2026. Writing the note on what each one does is your job in the video notes, so the notes are not here.

First-person testimony

• ASSUME THAT I CAN, CoorDown, World Down Syndrome Day 2024 (1:30, manual captions in seven languages).
• Q&A, Joshua and Sarah Littman, StoryCorps (4:00, manual EN and ES).
• The Saint of Dry Creek, StoryCorps (3:30, manual EN and ES).
• Terrie H., “Terrie, What Are You Doing?”, CDC Tips From Former Smokers (2:07, manual EN), with the 30-second cut of the same story (0:31). Watch the pair together.
• Traffic Stop, StoryCorps via PBS POV (3:19). Use the PBS link; the YouTube upload has no captions.

Explainer

• Housing Segregation and Redlining in America, NPR Code Switch (6:36, manual EN).
• Why is it so hard to escape poverty?, TED-Ed (4:45, manual captions in 34 languages).
• Slavery to Mass Incarceration, Equal Justice Initiative, narrated by Bryan Stevenson (5:50, manual EN).
• I had a black dog, his name was depression, WHO (4:18, manual captions in 16 languages).

Data-driven

• Wealth Inequality in America, politizane (6:23, auto captions only).

Observational and essayistic

• Why cities are full of uncomfortable benches, Vox (3:59, manual EN).
• The Long Wait, NYT Op-Docs, by Jason DaSilva (5:49, auto captions only).
• A Conversation With My Black Son, NYT Op-Docs (5:16, auto captions only).
• Look Beyond Borders, Amnesty International Poland (5:01, manual captions in six languages, auto in English).

Critique objects

• Most Shocking Second a Day, Save the Children UK (1:34, manual en-GB), and its sequel (1:56). Run the eleven unintended effects against both.
• Kids Respond to Child Hunger, SoulPancake (2:50, manual EN).

Caption audit

Of the twenty-two professionally produced advocacy videos checked for this list, four have only auto-generated English captions and one has none at all, and all of them were made by organizations with budgets and staff.

Video notes

Before Week 10, find 3 to 5 strong issue videos in your chosen form. Submit the list with one line on each: what works in its structure, its length, how it handles consent, what it actually asks the viewer to do, and its captions. Note explicitly whether each one’s captions are corrected or auto-generated. You bring the scan to the Week 10 lab, where it becomes the raw material for your brief.

Craft notes

Script length. Conversational English runs roughly 130 to 170 words per minute, and documentary narration is usually delivered slower than that. At 140 words per minute, three minutes is about 420 words and five minutes about 700, and you should cut ten to fifteen percent of that for pauses, breathing room over b-roll, and on-screen text beats. Most overlong scripts in this assignment are overlong by roughly that margin, which is enough to make the delivery sound rushed throughout.

Structure. Marshall Ganz’s public narrative moves Story of Self to Story of Us to Story of Now, which for this assignment has the practical effect of requiring the piece to move from a personal account to a collective claim to an ask. Kenn Adams’s Story Spine (once upon a time, every day, but one day, because of that, until finally, ever since then) is a quick test of whether the material you have is a story or a list.

Sourcing. Every number on screen comes from a named dataset and year, and you open the primary source itself before you use it. SIFT is the method, and lateral reading is the habit it is built on. Cite at point of use for any statistic that supports an argument, with a lower-third naming source and year, and put full references in an end-credit roll. Useful starting points: Census and ACS, BLS, KFF State Health Facts, Urban Institute data tools, HUD’s homelessness assessment reports, SAMHSA NSDUH, and CDC WONDER.

Archival and stock, with the licensing read properly. Free does not mean no attribution, and the rules differ per source. The Library of Congress’s Free to Use sets and the National Archives catalog are largely public domain and ask for credit as good practice. Prelinger has a per-item license and is roughly two-thirds public domain, so check the item and not the collection. Openverse aggregates CC and public-domain media and generates the attribution string for you, which is the workflow to use. Pexels requires no attribution but forbids showing identifiable people “in a bad light or in a way that is offensive,” which excludes using a stock face to depict addiction or homelessness. On music, Free Music Archive is per-track and its own warning is that free does not mean unconditional, and Incompetech is CC BY with prescribed credit wording. If your piece quotes news or archival material to make a point, the Code of Best Practices in Fair Use for Online Video describes the room you actually have, and a source log listing what you used and under what license is submitted with the piece.

Tools. Clipchamp ships with Windows and exports 1080p without a watermark on the free tier; DaVinci Resolve is free and far more capable if your machine can run it. Do not use CapCut for this assignment. Its terms grant a broad, perpetual, sublicensable license over uploaded content that extends to name, face, voice, and likeness, which is disqualifying for a course where you may be handling someone’s personal material.

Accessibility checklist

Accessibility is worth 10 points and is documented in the reflection. WCAG 2.1 makes captions on prerecorded media a Level A requirement, the lowest bar there is, and the W3C’s own position on auto-captions is that they do not meet accessibility requirements unless confirmed fully accurate, offering as evidence a machine transcript that turned “broil on high for 4 to 5 minutes, you should not preheat the oven” into “broil on high for 45 minutes, you should know to preheat the oven.”

Before you submit:

• Submit the corrected caption file itself (.srt or .vtt). A video with captions turned on does not satisfy this line.
• Full transcript submitted alongside the piece, describing what is on screen as well as what is said, which is what WCAG 1.2.3 accepts in place of a separate audio description track
• Captions follow the DCMP Captioning Key on the points that matter most: mixed case, speaker identification, and non-speech sound
• Text and text-over-image pass a contrast check (WebAIM Contrast Checker)
• Plain language throughout; a stranger outside the course can follow it the first time through
• Alt text on the thumbnail image and on any page where the video is hosted
• Nothing essential is conveyed by color or sound alone

All of this costs less at the storyboard stage than at export, so the storyboard template has accessibility as a column.

Milestones

Five-field reflection

The reflection is one structured form, the same for every video: objective and audience (the brief’s fields 2 through 5, and whether the finished piece still serves them), accessibility (what you did, per the checklist), ethical representation and standing (your positionality paragraph, the ethics standard you applied and how, consent and right of reply, voice and extraction, and which of Cho and Salmon’s eleven unintended effects your piece is most exposed to), AI use (disclosed and verified, with prompts logged for any generated asset), and what the other videos taught you (how the notes shaped what you made).

Rubric (100 points, scaled to 25%)

Fifteen of the hundred are awarded before the piece is finished, at the two points where a wrong turn is still inexpensive to correct.

Track and problem, due Week 5 (5 points)

Video notes, due the Tuesday of Week 11 (10 points)

The finished piece, due Week 12 (85 points)

FAQ

Can I pick any social problem? On Track B it has to be one you have seen firsthand, logged in your Problem Inventory, because that is where your standing comes from. If the problem you care most about is one you have only read about, come talk to me before Week 8; there is often a version of it you do have standing on. Track A starts from your own life instead, and Track C starts from the organization’s program.

Is Track A the easy one? No, and it is the one most often underestimated. It asks for less planning and more judgment: the 20-point ethics row is graded on how you handled the family, clients, colleagues, and supervisors who are in your story without having chosen to be, and a piece that solves that by leaving everyone out has to show in fields 5 and 7 that the omission was a decision rather than an oversight.

Which track should I pick? Track B unless you have a reason. Take Track A if your strongest material is about your own path and you are ready to handle the people inside that story. Take Track C if you already work with an organization that wants a piece, or if the Ginsberg archive holds footage on a subject you know something about.

Can I switch tracks? Up to the Week 10 deadline, yes. After that, the brief or the sheet is what the rest of the grading refers back to.

Do I have to put someone on camera? No, and for many topics you should not. The explainer, data-driven, observational, and essayistic forms all rely on archival material, graphics, objects, and narration. Whenever a real person does appear and is identifiable, the consent and right-of-reply rules apply to them.

Is this part of the Strategy Pitch? No. They are separate deliverables with separate rubrics and separate subjects. The pitch is a consultant memo arguing a technology decision; the video is a piece on the track you chose. If the organization you interviewed for the pitch is also the subject of a Track C video, that is allowed and it changes nothing about how either is graded, and neither one can be submitted as part of the other.

Can I make my video about an organization I work with? That is Track C. It needs the organization’s agreement to the brief before you shoot, and the consent and right-of-reply rules with everyone who appears. The video is not a course deliverable to the organization and is not sent to them by me. If you want to hand it over afterward, that is between you and them.

What is in the Ginsberg archive? Footage the Ginsberg Center holds from its own work with community organizations. It is available to this course by request through me rather than by browsing, so tell me the subject you are working on and I will determine what exists. Consent for the original filming does not automatically cover your edit, so anything you take from the archive is still subject to the consent section above.

Can I use AI-generated images or music? Yes, for backgrounds, textures, abstract motion, diagrams, and scoring, labeled inside the piece and prompt-logged in the reflection. Generated imagery may not depict real, identifiable people as if the footage were real, and it may not substitute for people affected by your problem, for the reasons in the ethics section. Generated shots also cost far more energy than text prompts, and the the energy and water note in the syllabus has the figures if you want them in the decision.

What if the person in my video pulls consent late? They can, up to the screening, and the rule exists to protect that option. The rough cut and right-of-reply viewing happen in Week 11 for that reason, which leaves three weeks of buffer to re-cut.

My topic is suicide, or sexual violence, or something similar. Is that allowed? Yes, with the topic-specific standard applied and cited, and come talk to me early. The reporting recommendations for these topics constrain real choices, including what you can show, what language you use, and what you put on screen at the end, so they have to shape the storyboard and cannot be checked at export.

How are pairs graded? Pairs receive one grade on one piece against one rubric, plus a short note on who did what. On Track B both people need some standing on the problem. Track A is one person’s story, so a pair there makes one person’s story with the other in a production role, named as such in the note.