Unpacking Uncertainty: The ABCs of Working with Real-World Data

articulate uncertainty interpret data to learn something

Real-world data is messy by nature. Values vary. Patterns aren’t always clear. Conclusions (always) come with caveats. Yet in classrooms, students often encounter data as if it’s clean and exact…giving an illusion of certainty that is far different than a reality we all know intuitively (we aren’t all the same height ;)).

Helping students learn to work with uncertainty in data—rather than avoid it—is a critical part of building strong data literacy. Recent work helping students make sense of uncertainty, including Rosenberg et al. (2025), offers some great ways to help our students realize that uncertainty isn’t a flaw in data, it is a natural reality of data. In fact, it’s a feature that, when made visible, can deepen student reasoning and scientific understanding.

In this post, we’ll unpack Rosenberg and colleagues’ ABCs of working with real-world data and explore how embracing uncertainty in data can support more authentic data sensemaking in classrooms this year.

Why Uncertainty Belongs in Data Learning

When uncertainty in data is hidden, students (and adults) often assume:

Data tells a single, definitive story
Graphs represent exact truth
Claims should be made quickly and confidently

In contrast, when uncertainty in data is made explicit, students begin to ask richer questions:

How confident are we in this result?
What might change the story?
What additional data would help us learn more about this system or situation?

These questions sit at the heart of scientific reasoning and data literacy. Rather than weakening conclusions, uncertainty strengthens thinking by encouraging students to weigh evidence, consider alternatives, and update their claims over time.

The ABCs of Working with Real-World Data

Rosenberg et al. (2025) introduce a helpful framework—the ABCs of data—to support how learners engage with uncertainty in data:

Acknowledge uncertainty

Build explanations with uncertainty in mind

Calibrate confidence over time

Together, these ideas shift data work from “finding the right answer” to reasoning with evidence while also deepening students' sensemaking in a particular data experience. Let’s explore more!

A: Acknowledge Uncertainty

The first step is simply naming that uncertainty in data exists.

In classroom data sets, uncertainty may come from:

Natural variability in systems
Sample size
Measurement tools or methods (note, this is the source most often presumed by students and is rarely a factor in quality datasets outside of the classroom…so be mindful of how you talk with students about it)

Instead of smoothing these away by averaging data or removing “outliers” from a dataset, let’s help our students identify and explore them. Asking questions like:

Which data points seem more or less reliable given what we know about the system?
What don’t we know yet?

Acknowledging uncertainty in data helps students understand that data represents observations of the world, not perfect replicas of it. It also helps to norm set that variability in our data is natural and to be expected. And that thinking about, rather than avoiding, the uncertainty in data will help us work with the data better.

B: Build Explanations That Account for Uncertainty

Once uncertainty in data is acknowledged, students can begin building explanations from the data that reflect it.

Rather than making absolute claims, students might learn to say:

“The data suggests…”
“Based on the available evidence…”
“This pattern appears likely, but…”

This language shift matters. It helps students:

Connect claims directly to evidence
Recognize the limits of what the data can support
Compare competing explanations fairly

In science, social studies, and even math contexts, this approach supports deeper reasoning and more productive classroom discussions.

Dataspire’s Advancing Pattern Recognition (grade banded) digital resource is designed to support exactly this kind of thinking—helping students move from noticing patterns to explaining them thoughtfully.

👉Explore the Advancing Pattern Recognition digital resources

C: Calibrate Confidence Using the Confidence Updater

One of the most powerful ideas from Rosenberg et al. (2025), is the notion of a Confidence Updater—the idea that confidence in a claim should change as new data or evidence becomes available.

Instead of treating conclusions as final, students learn to:

Revisit earlier claims
Adjust confidence levels
Explain why their thinking changed

This practice reinforces that learning with data is an iterative process, not a one-time event. And that is a HUGE mindset that students need to embody in order to successfully work with and make sense of data.

In classrooms, this might look like:

Revising written claims after new data is introduced
Comparing early and later interpretations
Reflecting on how confidence shifted over time
Using the Confidence Updater that Rosenberg and his colleagues put together

Why This Matters Across Grade Levels and Subjects

Uncertainty in data isn’t just a passing topic we acknowledge in Chapter 1 but then move on from—it’s a data concept that applies across disciplines.

In science, it is a reality in all real-world systems we investigate and in every dataset we look at
In social studies, it is fundamental to considering limitations in historical and demographic data
In math, it is necessary for students to reason about variability and distributions (specifically a part of middle school Common Core Math standards…but components are throughout the K-12 math experience)

By explicitly teaching students how to work with uncertainty in data, we help our students develop transferable skills that support critical thinking far beyond a single unit or course.

Supporting Teachers in Teaching Uncertainty in Data

Teaching uncertainty in data can feel uncomfortable at first—especially in classrooms where students expect clear answers. But with the right structures, it becomes an opportunity rather than a challenge.

Helpful strategies include:

Using real, imperfect datasets
Modeling in teacher talk how to think about uncertainty in data
Encouraging multiple interpretations of any dataset
Revisiting claims over time

Dataspire’s Advancing Pattern Recognition brings together tools and strategies that support these practices across grade bands, helping educators scaffold data reasoning in developmentally appropriate ways.

Final Thoughts: Uncertainty in Data as a Strength

When students learn that uncertainty in data is part of working with real-world data, they become more thoughtful analysts and more careful claim-makers. They learn to slow down, examine evidence, and adjust their thinking as new information emerges.

In a world saturated with data-based claims, these habits of mind are essential both within our classrooms and beyond. Embracing uncertainty doesn’t weaken learning—it deepens it.