What Does It Mean Today to Be Evidence-Based in Teaching? Part 3

CKHub: Making Professional Data Science Tools Work for Teaching

Introductory statistics is one of the most widely taken courses in higher education and often students’ last formal exposure to statistical reasoning. Yet many courses rarely involve working with real data, building models, or conducting reproducible analyses – despite data science making these central practices across professions.

This raises a fundamental question: if students live in a world shaped by data and computation, why don’t our statistics courses reflect that world?

The answer is rarely about instructors’ intentions. Most want students to engage authentically with data, but professional data analysis tools are often difficult to integrate into real classrooms – especially at scale and in equitable ways.

Being evidence-based in teaching isn’t just about what we teach. It’s about the learning environments we design.

Why “just adding data science” doesn’t work

Over the years, we’ve seen many attempts to “add” data science to statistics courses. Too often, this looks like bolting a tool onto an existing curriculum:

• Software students must install and configure
• Version conflicts across devices
• Limited instructional time lost to troubleshooting/IT
• Tools that work for some students, but quietly exclude others

When computation becomes fragile, learning becomes fragile too. Instructors end up spending more time managing tools than supporting thinking. Students who hit technical barriers disengage – not because the ideas are beyond them, but because the environment wasn’t built for them. If we want data science to be part of introductory statistics in a meaningful way, the tools must meet classrooms where they are.

Evidence-based tools must make learning possible at scale

At CourseKata, we think about tools as instructional infrastructure. CKHub was designed around three core principles that make professional data science tools work for teaching.

1. Make computation easy

In statistics courses, the challenge should be reasoning with data – not installing software. CKHub removes technical setup as a barrier. There is no software to install. Everything runs directly in the learning management system. Instructors can partially write code, scaffold key steps, and focus students’ attention on interpretation rather than syntax. When setup barriers disappear, a broader range of students can participate fully. Class time can be spent on ideas, not logistics. Teachers can teach.

2. Make work reproducible

Reproducibility is not just a technical ideal – it is a social learning practice. In CKHub, students can run their own analyses, run each other’s work, and revisit their thinking over time. Instructors can run student notebooks to understand how conclusions were reached. Work can be shared, revised, and discussed. This makes learning visible – not just final answers, but the reasoning and decisions that produced them. Reproducibility supports accountability, collaboration, and reflection. It turns student work into something that can be examined and improved, rather than something that disappears after submission.

3. Make analysis coherent

Computation alone is not enough. Statistics is about telling meaningful stories with data. Professional data scientists don’t just write code – they explain what they did, why they did it, and what others should notice. Jupyter notebooks were invented to solve exactly this problem: integrating computation, explanation, and interpretation into a single, coherent narrative. CKHub brings this same structure into the classroom. Code, text, and visualizations live together. Students are guided to explain their choices, interpret results, and connect statistical output to real questions. This coherence is essential for learning. It helps students move beyond “running analyses” toward thinking with data. It also aligns directly with the kinds of reasoning we want students to transfer beyond the course.

Tools as pedagogy

CKHub isn’t evidence-based because it uses data or code. It’s evidence-based because it supports environments where meaningful learning can happen – environments where:

• Students engage with real data
• Reasoning is visible and revisitable
• Teachers can see and respond to student thinking
• Computation supports understanding instead of overshadowing it

Modernizing statistics education isn’t about adding more technology. It’s about designing tools that align with what we know about learning and teaching. If introductory statistics is going to prepare students for a data-driven world, then computation, modeling, and reproducibility can’t be optional extras. They have to be built into the everyday work of the classroom –in ways that are inclusive, coherent, and sustainable.

That’s the work CKHub was designed to support. 

Watch this 18-second demo to see CKHub in action:https://drive.google.com/file/d/1ftjoCIxaY4i7Orm4x2kHbrl8p0wphkAs/view?usp=sharing