Who Gets Counted? Insights from the 2026 ChangeInSight Report

Data shapes how communities are seen—and just as importantly, what goes unseen. Our 2026 ChangeInSight report shows that for AANHPI communities, the way data is collected and grouped can blur critical differences, real needs and lived experiences.

These gaps show up in different ways: overlooked subgroups like the Chin community, limited language access, and inequities in how rees are distributed. Across each finding is the same pattern—when data is too broad or incomplete, entire communities fall out of the picture.

Here are five takeaways that highlight why disaggregated, community-informed data isn’t just helpful—it’s necessary.

When AANHPI communities are grouped into a single category, important differences between subgroups disappear—making it harder to identify who needs support and how.

Our report highlights the Chin community as an example. While often categorized as “Burmese” or “Asian,” the Chin are a distinct ethnic group from Myanmar with unique linguistic, cultural and religious identities. Many arrived in the United States as refugees, shaping experiences that differ from the broader Burmese population.

Language is one clear example. Hakha Chin—not Burmese—is widely spoken in the community. When data groups Chin individuals under a broader identity, their specific language needs are overlooked, leading to gaps in access and support.

Incomplete data doesn’t just limit understanding—it directly affects how resources are distributed.

Our surveys found that 14.6% of respondents face transportation barriers, but that number alone doesn’t tell the full story. Disaggregated data reveals a 22% gap between Filipino and Vietnamese respondents, with nearly 33% of Vietnamese respondents reporting barriers compared to 11.2% of Filipinos.

These differences are often lost in national datasets. Without this level of detail, it becomes harder for policymakers and service providers to identify which communities face the greatest challenges—and to allocate funding accordingly.

Another key issue with national datasets is that they are built without meaningful input from the communities they represent, leading to incomplete or inaccurate conclusions. Rather than tracking trends, ChangeInSight’s research focuses on the experiences and needs of local populations so that community organizations can better serve their communities.

Such niche data is important because local organizations often lack the resources, technical support and infrastructure to consistently collect and interpret data.

Language access is often treated as a service issue, but it starts much earlier—with how data is collected.

When surveys are English-only or offer limited language options, entire communities are excluded from participating. As a result, the data itself becomes incomplete.

While federal datasets may include 20–25 Asian ethnic categories, many policy efforts rely on even fewer. ChangeInSight’s research includes over 85 ethnicities—many of which are typically grouped into “other Asian.” This includes communities like Cambodian, Laotian and Chin, whose needs often go unrecognized when they are not explicitly represented.

Data equity is often misunderstood as simply collecting more data—but more isn’t always better.

What matters is whether data is representative, accurate and shaped by the communities it reflects. Without that, even large datasets can reinforce the same gaps and inequities.

True data equity means moving beyond quantity and focusing on quality—ensuring that communities are not just counted but meaningfully included.

At its core, data equity isn’t just about improving numbers—it’s about improving visibility. When communities are flattened into broad categories or excluded altogether, the result isn’t neutral—it shapes who gets resources, who gets recognized and who gets left behind.

What this report makes clear is that better data starts with better listening. When communities are involved in how data is collected, defined and interpreted, the result is not only more accurate—it’s more actionable.

If we want policies and services that actually reflect the people they’re meant to support, we can’t rely on incomplete pictures. Data should do more than count communities—it should represent them.