from Narrative Observatory
Jan 2023

Health Equity Narratives: Content Testing & Strategy Validation

Our flagship research from 2022 to study audiences and the health equity narratives that move them with generous support from Robert Wood Johnson Foundation and in partnership with Story Strategy Group.


Storytelling for health equity

As the nation’s largest philanthropy dedicated solely to health, Robert Wood Johnson Foundation (RWJF) has been working to help achieve health equity with transformed healthcare and public health systems that are fair and just. In 2021, RWJF enlisted our support to build an audience-narrative architecture that allows RWJF and its partners to develop content that can change the narrative on health equity. You can explore our findings in the full report.

Here, we focus on the third and final phase of the project—content testing and strategy validation. In this phase, we built stories that could actually move audiences in the right direction and evolve their narrative landscape, turning stories into ads that could be tested for engagement and transportation.

Table of Contents

Strive & CreateUs
Community IconPeople powerValue: Community
Autonomy IconIf you say soValue: Autonomy
Order IconTough cookiesValue: Order
Authority IconDon’t tread on meValue: Authority
MeProtect & Preserve
Learn more on our values-based audience segmentation here

Creating an audience-narrative architecture for changing the narrative on health equity

We started with our 4 values-based audience as a means of partitioning stories. Then found the stories about health and healthcare that are distinctive to each audience and suggest a narrative landscape on health and equity. Finally, we used survey data to identify key attitudinal differences in our how 4 audiences think about health, healthcare, and health equity in the U.S.

4 audiences with 4 unique health equity narratives

Next we explored each audience’s narrative landscape. Here, we started with all the stories our 4 audiences consumed across online news, TV news, entertainment TV, and Twitter, between January 2021 and December 2021. What emerged were four narratives, or one distinct narrative per audience. Explore more about these four narratives in the full report.

  • People Power

    Everyone needs to play their part to keep healthy—physically and mentally—mostly by following the advice of medical experts.

  • If you say so
    trust + Care

    It’s hard to know who has the right answers when it comes to health. Getting care means navigating obstacles and finding allies.

  • Tough cookies

    The powerful play an important role protecting us from disease, as we must protect others, especially children.

  • Don’t tread on me

    The primary vector for ill health isn’t disease; it’s other humans, from politicians to immigrants, scheming to injure or infect us.

Our narrative findings were supplemented with qualitative research wherein we interviewed and solicited imagery from eight people for each of four audiences to collect their personal stories on health. We used AI to analyze these user-submitted images to generate prompts that were then fed into Stable Diffusion AI to generate the videos above.

Each narrative had a different tone, different protagonists, different challenges, different ways of viewing race and racial disparities in health and healthcare. These differences in narrative landscape informed our hypotheses about the content most likely to resonate with audiences and evolve their narratives on health equity.

01. Testing Program Design

Before we start experimenting with building stories, we need a target narrative to guide and inspire the stories we tell.

With the audience-narrative architecture in place, and in collaboration with our partners, advisors, and RWJF’s Health Equity Collective we developed a target narrative that could help measure success.

Target narrative

“A new system is needed to replace our current, racially biased healthcare system: one that is accessible and affordable to all, delivers excellent health outcomes across all populations, recognizes and rectifies past and current injustices, and honors everyone who participates in it.”

Generating hypotheses

Over the course of six months, together with project partners, we used a few different methods to generate hypotheses about the stories that might move each audience closer to our target narrative.

  • 1
    Media analysis

    We re-analyzed each audience’s media diet for content that might already be moving them toward our target narrative or holding them back. We called these bridge and barrier hypotheses and generated 4 per audience.

  • 2
    Subject matter experts

    Our project partners convened professional storytellers, anti-racism experts, community leaders, and content strategists to collaboratively generate 13 liberatory hypotheses, embodied in short stories.

  • 3
    Creative strategists

    A team of professional creative strategists synthesized learning from the first two rounds and also from existing supplemental research.

From hypotheses to content: the lifecycle of an ad

In total, across all three testing phases, we created 103 pieces of content, or “ads,” embodying 56 different stories. Our approach to testing is iterative, meaning stories that originated in one phase evolve in subsequent phases. For example, here is an ad that originated in round 2 and was expanded upon for three different audiences in round 3:

Click on an ad to see the accompanying social media headline and copy or click here to explore all 103 ads.

From liberatory hypotheses to immigrant healthcare worker ad

Liberatory hypothesis #7: Everyone deserves quality healthcare, regardless of legal status. This liberatory hypothesis was transformed into an ad intended to reach all four audiences.

When testing content, both engagement and transportation matter

When we test content to see how audiences respond, we care about engagement and transportation. Engagement is whether audiences view, click on, or share the content, when it appears inside a social media feed. Transportation is the extent to which content transports or moves audiences toward a target narrative. This can also be referred to as persuasion. We test this using randomized controlled trials via a custom-built survey platform and these three survey prompts.

  • A problem exists

    “In today's healthcare system, regardless of their income, white people generally get better care than people of other races.”

  • It should change

    “Right now, white people have better healthcare outcomes than people of other races. We should change our healthcare system to give everyone the healthcare they need to stay well, even if it means that my family and I have to give something up.”

  • We should change it

    “I can imagine how we could change American healthcare in the next 5 years, so that patients of all races feel respected and have access to the care they need and want.”

Measuring baseline agreement

Our survey questions for transportation testing arose from a review of existing survey research and in consultation with our project partners, including equity advisors. The questions needed both to capture movement toward our target narrative, and also provide audiences enough room to move. In other words, we were looking for low to moderate baseline agreement, before treated audiences viewed the content we created.

A problem existsIt should changeWe should change it
Pre-testing baseline agreement for each audience across the 3 survey questions.
03. key Insights

A sample of standout stories

Of all the stories we created and tested, there were some notable standouts that consistently performed well in engagement testing, transportation testing, or both. These can provide important direction for future iteration.

This story A was the only one to transport all four audiences, with all ads performing well across both engagement and persuasion. Themes from this story like respect, control, trust, and community, might be a reason for its universal success.
RWJFround2_story1b (2).jpg
RWJFround2_story1c (3).jpg
Two ads from the “Midwife” story caused backlash for People Power in Round 2 (A and B), but this ad Cfrom the “Birthworkers” story moved them towards the target narrative, a testament to how iterative testing can bring an audience around.
Two different ads from the “Birthworkers” story also performed surprisingly well with Don’t Tread on Me Aand B, although they were designed specifically for People Power.
IYSS 1A. Coaches.png
DTOM 1B. Coaches.png
While many ads for the “Coaches” story worked with If You Say So A, this one B designed for Don’t Tread on Me and this one C designed for People Power caused backlash. Stories are not “one size fits all.”
This ad A, designed for all audiences, did well at moving If You Say So. A later iteration B, was designed specifically for If You Say So but ultimately was highly engaging for all audiences
key Insight #1

When we tested content for transportation, baseline agreement affected how far audiences moved

In our transportation testing, whenever there were differences by race in how audiences responded to the content we tested, white people almost always moved more, because they were starting at a lower agreement baseline than other racial groups.

One of the stories A we tested, which centers disparities faced by affluent, college-educated Black people, moved people of color across audiences more than white people.
Occasionally, a story moved white people, but caused backlash for people of color in the same audience. This happened with both the “Coaches” adA for Tough Cookies, and the “Superhero” ad B for If You Say So. People of color may be more onboard with the target narrative generally, but it is possible to create on-narrative content that pushes them away.
The “Palette” story expressed like this A worked for every audience but Don’t Tread on Me. The same story expressed like this B did work for Don’t Tread on Me. This was the most successful expression of this story in most audiences, and scored high on engagement overall. This suggests that content featuring people outperforms the more conceptual. It also suggests that audiences can move explicitly on race, even when racial difference is implicit, versus stated explicitly.
key Insight #2

Formats with futures make for a better story

On average, content created as a diptych—two panels of text and images—performed better than single-panel images. In single panels, there is no room for development or resolution. A second panel can help transform, channel, or address whatever the first panel might bring up in audiences. Diptychs afford more specificity—more of the implied time, place, and situation that makes for good story. And story works well across audiences.

Content Warning. Click To View
The “Arrested” story A worked in People Power and in Don’t Tread on Me, with both diptychs (A and B) producing movement.
We saw this again in how audiences responded to the “Birthworkers” story. It was the diptych A, not the more poignant problem-focused imagery B, that created more engagement across audiences.
key Insight #3

Aim for the edges to find bridging opportunities between audiences

Engaging and transporting new audiences outside of our existing core audience is key to achieving the target narrative. Speaking to the existing base is relatively easy—they’re already “on board” with the target narrative. Aiming for the edges of that base audience is where new and untapped transformation can happen.

The Native American “Diabetes” story A was the only one to transport all four audiences, effectively working in both engagement and transportation testing. Its themes of community voices, respect, trust, and control may account for its success. It’s also a very specific and personal story, located in a specific place and a time in history.
key Insight #4

Unintended consequences: in some instances audiences were transported in the wrong direction

Some ads caused backlash within specific audiences. These instances do not indicate that the story is ineffective, but offer opportunities to explore why an ad might not have represented the original story well enough and provide insight for different approaches.

DTOM 1B. Coaches.png
While If You Say So moved on multiple ads in the “Coaching” story, two ads caused backlash: “Coaching” A designed for Don’t Tread on Me and “Coaching” B designed for People Power.
Content Warning. Click To View
The single “Arrested” A ad designed for all audiences caused backlash in Tough Cookies.
The single “Birthworkers” ad A designed for Tough Cookies caused backlash in People Power. This might signal the importance of utilizing diptychs to tell stories.
04. audience insights

Audience upshot

As practitioners in narrative change work, it is important—necessary, even—to think about and aim for audiences who have values that may not be familiar to us. The narrative architecture built in this work intended to teach us more about what really motivates audiences, allowing for the creation of engaging, effective content. Below, we share audience-specific insights.

People Power is a strong base for the narrative goal

Community and universalism drive People Power: understanding, appreciation, tolerance, and protection, for the welfare of all people and for nature. In other words, they really care about people they may not know, and their Medical narrative reinforces an attention to inequitable systems. This is the audience most likely to engage with content centering systemic racism. Importantly, the bulk of content that transports other audiences also transports People Power, too. By aiming creative efforts outside of People Power, it may be easier to build broad-based narrative power.

Screenshot 2023-01-22 at 3.15.21 PM.png
Ads that work with other audiences often work with People Power, like the “Bad Blood” B ad designed for Don’t Tread on Me.

Content with a clear future is critical to moving If You Say So, and speaking their unique cultural language is key to engagement

If You Say So is less willing to engage than other audiences generally. The content that does engage tends to be sillier, darker, and more cynical than the stuff that lifts heart-and-home Tough Cookies, for example.

IYSS 3B. Confront Racism.png
Silly and cynical meme-like creative engages If You Say So, like this ad A designed specifically for them.
Most of what engages features individuals, not families or groups, like these two ads, “Arrested” A and “College” B.
IYSS 1A. Coaches.png
DTOM 1C. Coaches.png
Three different versions of the “Coaches” ad, A designed for If You Say So A , B designed for Tough Cookies B, and C designed for Don’t Tread on Me, all moved If You Say So. This is consistent with their Trust + Care narrative, in which they’re looking for trusted allies to help navigate healthcare systems.

Tough Cookies: easy to engage but harder to persuade

Tough Cookies gravitates toward content featuring heart and home, faith and family, plus other types of community working together. Tough Cookies was especially moved by content about changing the system, for example, to allow providers trained outside the U.S. to practice, rather than work as drivers.

IYSS 4A. Jean-Pierre.png
Multiple ads in the "Health Credentials" story worked with Tough Cookies, including these designed for If You Say So A & B, and this one designed for People Power C.

Effective Don’t Tread on Me content centers control and choice.

We were doubtful of being able to reach and move Don’t Tread on Me, given the antagonism of the Enemy narrative. We found that the antidote to this narrative is control and choice. All the content from the “Sick” story moved Don’t Tread on Me, given its focus on personal control.

IYSS 2A. Sick Day.png
Though designed for other audiences (If You Say So A and Tough Cookies B), two ads from this story worked with Don’t Tread on Me.

The average effects tell a story about how each audience’s specific values can contribute to achieving a new narrative.

The average effects we observed in audiences from treating them with content correlated to the values and health-related narratives which define each audience. They also point to the important contributions each audience may be able to make to achieving a new narrative on health equity.

  • peoplePower
    People Power easily recognizes the problem (1), the need for change (2), and pathways for realizing change (3).
  • ifYouSaySo
    Creative, autonomous If You Say So moves most toward imagining the future (3).
  • toughCookies
    Safety-minded Tough Cookies moves most to perceiving the threat (1).
  • dontTreadOnMe
    Action-oriented and leadership-driven Don't Tread On Me moves most toward demanding change (2).
A problem existsIt should changeWe should change it
The average effects we observed in audiences from treating them with content correlated to the values and health-related narratives which define each audience.

Each audience has different entry points to the target narrative

It is possible to align all four audiences to a narrative on health equity that recognizes the problems inherent in our current, racially biased health system, and posits a new system that rectifies injustices, honors everyone who participates in it, delivers excellent health outcomes, and is accessible and affordable for all populations. But each audience approaches this narrative from different places and perspectives. Overall, lifting up choice and control, in addition to injustice and outcomes, allows us to reach and move individually focused audiences also.

Screenshot 2023-01-22 at 3.15.21 PM.png
People Power responds the most to direct appeals for justice, equity, and fairness in the content they engage with A. They easily recognize the problem, the need for change, and pathways for realizing change.
IYSS 1A. Coaches.png
If You Say So responds to content that reflects a desire to better navigate the system to satisfy their needs A. Creative and autonomous, they move most toward imagining the future.
IYSS 4A. Jean-Pierre.png
Tough Cookies responds to content that frames change via better outcomes in their communities and families A. Safety-minded, they move the most to perceiving the threat.
DTOM 2C. Sick Day.png
Don't Tread On Me responds to content that promises more personal control and choice, for themselves as part of everyone A. Action-oriented and leadership-driven, they move most toward demanding change.
05. Conclusion

From insights to action, general tips and takeaways.

  • 1

    Put People In It

    Content featuring people outperformed content that did not, and photography tended to outperform illustration.

  • 2

    Get Specific

    Content that established a specific time, place, situation, characters, and some kind of development performed better than more conceptual content. Good story-telling wins again!

  • 3

    Tell a story with a future

    In general, the diptychs we tested outperformed single-panel ads. Finding a structure to name the problem and the desired future is key.

  • 4

    Always be iterating

    The best performing content rarely emerged in our first attempt but, rather, over many rounds and iterations, with input from many people. We encourage you to explore all the ads we made and iterate on them.

Strive & CreateUs
MeProtect & Preserve
These maps show ads with high-ranking engagement or transportation. The position of the content on the map shows its “home base” for maximum engagement or persuasion.
Bridging opportunities

We can look to story features as clues about what could work to transport audiences in the future

Because we tested all the content we created across all four audiences, even if we were targeting just one audience, we were able to learn about both the content that distinctively engages or transports a single audience, and the content that appeals across audiences. This visualization shows content ranked by engagement across all four audiences.

More from the Narrative Observatory @Harmony Labs

Researching and reshaping our relationship
with media

In partnership with advocacy organizations, philanthropic foundations, entertainment and news media brands, and creators, the Narrative Observatory has helped shape the cultural strategy and media making of dozens of partners to-date. You can explore some of this work below:

If you’re interested in learning more about the Narrative Observatory and how it might be leveraged for your work, get in touch.