Teaching

My teaching is grounded in the belief that students learn best when they are supported through productive struggle, clear expectations, and meaningful opportunities to practice. I am especially interested in how students learn technical skills in the AI era while developing agency, persistence, transparency, and critical judgment.

DATA-R155 redesign

I teach the foundations of the statistical programming language R in DATA-R155: Introduction to R Programming. This is a one-credit course at Boise State University.

Generative AI has made R code easier to produce. Beginning programmers still need ways to understand what the code is doing, determine whether it worked, and explain how it connects a question to a result.

Since summer 2025 semester, I’ve iteratively redesigned DATA-R155 around a question: How can an introductory programming course help all students make their thinking visible as they learn to work with R and generative AI? My answer was to the learning process part of the documented workflow. The course asks students to explain code, document how their understanding develops, and practice using AI in specific guided ways.

The redesign takes form through two connected instructional resources. A digital open educational resource guides students through R, learning routines, and selected AI practices. A Canvas course environment organizes pacing, assessment, student choice, and submitted work. I designed, authored, built, and teach these resources with instructional-design guidance from Boise State eCampus and professional-development support from the Center for Teaching and Learning.

Open educational resource

I am designing a living open educational resource (OER). This textbook is publicly available on GitHub and is designed to be iteratively improved based on student feedback and my own teaching experience.

Explainable and reproducible work

The goal is to teach students how to learn R, covering more breadth than depth. Most R programming skills are introduced at the “understanding” level in Bloom’s taxonomy. Students are expected to explain how code works; memorizing functions is not a course expectation. The OER also showcases selected RStudio capabilities, providing students with ideas on how learning R can be extended beyond the course.

The text does not only introduce R navigation, syntax, debugging, data manipulation, and visualization at an introductory level. It’s activities teach formatting for an R script, R Notebook, and Quarto files that place code within a written account of learning reflections. I emphasize reading and explaining code, investigating unfamiliar functions, and leaving a record that makes the work easier to revisit and follow. Students learn how to format their R files into sharable HTML reports.

The design is intended to empower students so that they can independently learn the techniques for their data analytic needs after the course. This approach makes the course applicable to students with a wide range of backgrounds and future goals.

Self-regulated learning routines

Novice programmers need explicit support for problem-solving processes (Loksa et al., 2016). Students learn how to manage confusion and strengthen their metacognitive awareness with the aid of recurring emojis 🎯 Goal, ⚡ Code, ✅ Verify, and 🗣 Reflect. Emojis are used to illuminate the routine.

Learning skills embedded

Along with the learning routine emojis, I use emojis to highlight learning skills that support the development of intellectual virtues (Baehr, 2022). Each module foregrounds a learning skill: 💪 tenacity, ✍️ transparency, 🦁 courage, 🧭 autonomy, 🤔 open-mindedness, 🐱 curiosity, or ✔️ thoroughness. Using the emoji, I model how to document the highlighted skill within the guided activity. Then I prompt students to practice the skill within the practice space.

Because the online OER is still being developed and used in an active course, the public chapters change as modules are revised or released to students.

Canvas learning environment

The Canvas site is the working course environment students use to access modules, complete activities, submit work, and receive feedback. It links the OER to low-stakes core assessments and advanced learning activities.

To support an alignment with outcomes, promote revisions, and encourage students to take ownership of their learning, I’ve carefully designed the assessment structure.

Specifications grading

Research on feedback and assessment shaped this choice. Self-regulated learning routines promote the use of feedback to revise work, but grades can redirect students’ attention away from comments and toward feelings of judgment. The presence and value of a numerical grade can also shape how students perceive written feedback (Black et al., 2004; Lancaster et al., 2020). Because mastery-oriented assessment can change the meaning of assessment (Smeding et al., 2013), I selected specifications grading to create conditions for self-regulated learning. This grading system offers one structure for aligning assessment with mastery, feedback use, and revision (Nilson & Packowski, 2025).

This approach organizes the course around work evaluated as ✅ Complete or 🟡 Not Ready. All students complete a shared 🌱 core to earn a passing grade (C), then select from a larger set of 🌳 advanced activities to earn a higher grade (B/A). Completing activities unlocks more activity choices, allowing students to decide where they want to extend their learning and what work they will undertake toward their intended grade.

AI as a resource for learning

LLMs may interfere with novice self-regulation when students move directly from prompt to generated solution, bypassing planning and reflection (Prasad & Sane, 2024).

I designed guided activities that help students use generative AI to investigate code and test how well they can explain it. Students are offered several custom course chatbots that can answer questions about R, the course, and provide code as part of specific guided activities. Additionally, I created a few custom bots for myself to use for feedback and student communication.

Many activities offer optional AI-use guidance. For example, one activity asks students to explain a provided R script. Students can choose whether an AI tool or a classmate will reconstruct the code from that explanation. The reconstruction makes ambiguities in the explanation visible. Similar choices appear throughout the course, giving students a choice between working with a peer or an AI system.

Instructional design

I designed, authored, built, and teach the course with instructional-design guidance from Boise State eCampus and professional-development support from the Center for Teaching and Learning. I also participated in Boise State’s faculty learning communities “Designing for Student Success” and “Accessible by Design,” applying that work to the structure and presentation of DATA-R155.

The Canvas course environment offers multiple ways for students to navigate the course, including through a website-like page design, a module-based structure, and a weekly calendar. A consistent weekly structure helps students know what to expect and how to plan their work. I use Canvas Learning Mastery Gradebook, custom letter grade, and module requirements to support specifications grading and assignment access. The Canvas sandbox can be used for either online or in-person course delivery.

This redesign is one place where my teaching and research meet.

Scholarship of Teaching and Learning

The associated Scholarship of Teaching and Learning study will examine students’ experiences with the design and how it relates to course performance, later enrollment, and the learning practices visible in their work. The study asks how student success outcomes and equity patterns in redesigned DATA-R155 offerings compare with historical DATA-R155 offerings and a comparison course, and to what extent students continue into the R Certificate or subsequent R coursework.

Specifications grading is meant to create conditions for better learning behavior, but it does not guarantee it. Students may still focus on grades, resist the lack of partial credit, complete revisions superficially, or experience anxiety when expectations are unfamiliar. Grade-pathway choices, assignment completion, revision records, and selected reflections will help me examine how students navigate mastery, feedback, productive struggle, and course expectations.

In an AI-assisted introductory R course, the need for process evidence becomes sharper because final code output alone is too weak as evidence of learning. A runnable script, correct table, or polished report may show that a student reached an answer, but it may not show whether the student understood the code, checked the output, used AI appropriately, incorporated feedback, or revised a flawed workflow. The study will examine changes in students’ readiness for further R coursework and look for evidence that they are developing responsible AI-use judgment and a reproducible and explainable R workflow as defined in the course.

I presented the study design in the poster Redesigning DATA-R155 for the AI Era at Boise State’s 2026 Great Ideas for Teaching and Learning Summit.

Teaching certificates

My course development has been supported by continuing work in online teaching, instructional design, accessibility, and reflective course revision.

  • eCampus Center Intermediate Online Teaching, Boise State University, February 2026

  • eCampus Center Basic Online Teaching, Boise State University, November 2024

  • Teaching and Learning in Higher Education, Boise State University, August 2023