Large Lecture Courses

Overview

Teaching at scale is a specialized form of instruction. There is a specific kind of "stage fright" that comes with a large enrollment course; the fear of the "void", that 50-minute stretch where you’re talking to a room where many students are hidden behind screens. In a 300-seat hall, the laws of classroom physics change. Sound travels differently, the "back-row effect" creates a vacuum of attention, and the psychological distance between the podium and the student can feel insurmountable. And it can turn even the most brilliant scholar into a "talking textbook." To fix this, we have to move beyond thinking of the lecture as a "speech" and start seeing it as a choreographed social experience.

Strategic Course Design

Teaching always begins long before you step onto the podium. Whether you are leading a small class or a massive lecture, the most robust starting point is Backward Design (Wiggins & McTighe, 2005). This framework shifts the focus from "what I am going to say" (content-centered) to "what the students will be able to do" (learning-centered).

While Backward Design is the standard for all course development, applying it to a large-scale environment might look a little different. In a small class, you can often sense when students are lost and pivot in real-time. In a large lecture, we should build that "alignment" into the very bones of the course to ensure that 300+ students stay on the same intellectual path.

Stage One: Identifying Desired Results (Goal Setting)

The first stage of Backward Design is identifying what students should know, understand, and be able to do by the end of the term. In a large course, this isn't just a planning exercise; it’s about deciding what is most important for students to learn In a seminar, you can afford to explore the side topics of a discipline. In a large lecture, you must distinguish between what is "nice to know" and what is "essential to master." So, you have to be an editor of your own expertise to prevent "instructional drift.”

• Threshold Concepts: These are the high-level ideas that form the "heart" of your discipline, the (Meyer & Land, 2003). The authors define threshold concepts as transformative (they change how students see the field), troublesome (they may feel confusing at first), integrative (they reveal connections between ideas that previously seemed unrelated or hidden), and often irreversible once understood. These can serve as anchors for lecture goals.
• Important to Know and Do (Skills): These are the essential skills and methods (e.g., how to analyze a primary source or solve a specific equation). These goals are often best reinforced in discussion sections or through asynchronous practice.
• Worth Being Familiar With: This is the broader survey information. At scale, this material is ideally moved to supplemental readings or pre-class materials, ensuring it doesn't dilute the focus of your limited synchronous sessions.

Stage Two: Determining Evidence (Assessment at Scale)

In any teaching context, this phase is where we define the metrics for success: “How will I know if they’ve actually met the goals I set in Stage One?” In a high-enrollment course, this stage becomes a logistical and pedagogical "stress test." The challenge is to design a system that provides meaningful evidence of learning for hundreds of students without exceeding the instructional team's capacity to provide timely, actionable feedback.

Testing Frequency

A common pitfall in large courses is the "Midterm-Final" trap; two high-stakes events that provide feedback far too late for a student to course-correct. Shifting toward high-frequency, low-stakes assessments helps students adjust their learning earlier and improves long-term retention (Roediger & Karpicke, 2006; Dunlosky et al., 2013). Monthly or bi-weekly automatically graded "Learning Checks" on bCourses serve two purposes:

1. They force students to engage with the material continuously.
2. They provide you with a real-time dashboard of student understanding. If 60% of the class misses a specific quiz question, you can dedicate a session to that specific "bottleneck."

Designing for Learning at Scale

The primary challenge at scale is translating high-level academic standards into a system that remains fair, consistent, and manageable across a large cohort and a diverse teaching team. See examples below:

Promoting Integrity through Design

In a room of 300, surveillance-based proctoring is often less effective than integrity-based design. Research on academic dishonesty shows that students are more likely to engage in misconduct in high-pressure, high-stakes environments, particularly when expectations are unclear or perceived as unfair (Krou et al., 2021; Lang, 2013; Gallant & Rettinger, 2025).

Rather than relying solely on detection, many instructors design assessments that reduce both the opportunity and the motivation to cheat. This approach shifts the focus from policing behavior to structuring conditions that support integrity.

Clarify Expectations Early and Often

Clear, repeated communication of expectations plays a critical role in supporting academic integrity. Students are more likely to act with integrity when norms are explicit and reinforced.

• Remind students of Berkeley honor code and course-specific expectations.
• Include a brief integrity affirmation at the start of an exam (e.g., a written pledge).
• Clearly articulate your policy on AI use. Ambiguity around tools like generative AI can increase confusion about what is permitted. As needed, remind students of the community standards for integrity at the assignment and activity-specific levels.

Design Assessments That Reduce Incentives to Cheat

Academic integrity is closely tied to assessment design. When tasks emphasize recall under pressure, students may be more likely to seek external assistance. When tasks emphasize reasoning, application, and course-specific thinking, shortcuts become less useful.

• Emphasize reasoning and explanation. Ask students to justify their answers, interpret results, or connect concepts to course-specific materials (e.g., lecture examples, datasets, or case studies).
• Test your own prompts using generative AI tools. If an AI system can easily produce a correct answer, consider revising the question to require deeper engagement.
• Reduce unnecessary competitive pressure. Curved grading systems or environments that emphasize ranking may increase anxiety and the likelihood of academic dishonesty (Krou et al., 2021; Lang, 2013). Transparent grading criteria and rubrics can mitigate this effect.
• Consider open-resource formats. Open-book or open-note exams can shift the focus from restricting access to evaluating how students think, analyze, and make decisions.

Use Structure and Technology Strategically

No technical solution can fully prevent access to external resources. However, thoughtful use of course tools can reduce opportunities for misuse while maintaining accessibility.

• Use time limits on bCourses quizzes
• Display one question at a time or limit backtracking when appropriate.
• Be mindful of accessibility needs; students with accommodations may require extended time or alternative formats.

Reinforce Integrity as a Shared Value

In large courses, students can feel anonymous and disconnected, which may weaken their sense of accountability. Framing integrity as a shared responsibility within a learning community can help counter this effect.

• Explain why integrity matters in your discipline; not just as a rule, but as part of scholarly practice.
• Signal trust while setting clear boundaries. Explain your expectations in a way that emphasizes learning rather than suspicion. For example, instead of saying “cheating will not be tolerated”, maybe consider: “this assignment is designed to help you practice [SKILL]. You’re welcome to use [X] but not [Y]. If you’re unsure, please ask.”
• When possible, design follow-up opportunities (e.g., brief oral explanations or reflections) that reinforce accountability for submitted work.

Stage Three: Planning Learning Experiences

Stage Three of Backward Design is planning the learning experiencesIn a standard course, this is where we design the day-to-day activities. The fundamental shift for a large course is moving from pushing information out to helping students process the information. Because you have already identified your "Threshold Concepts" in Stage One and set up your assessment in Stage Two, the lecture now serves a specific, high-value purpose: resolving the most difficult concepts in real-time.

Segmenting Lecture

Attention fluctuates over the course of a lecture (Johnstone & Percival, 1976; Bunce et al., 2010; Szpunar et al., 2013). To maintain engagement at scale, the instruction can be designed in "waves" of delivery followed by processing.

Considerdividing your 50- or 80-minute block into segments. After each segment of direct instruction, insert a cognitive reset. This is not necessarily a break, but a brief, targeted activity that forces students to move information from working memory to long-term storage.

• Free Recall: Ask students to put away their notes and write down the most important concept from the previous 15 minutes.
• The "Muddiest Point": Ask students to circle the one concept that feels the most "muddy" or confusing from the notes. Have them attempt to explain it to a neighbor or formulate a specific question about it.
• Note Comparison: Give students 2 minutes to compare notes with a peer to identify gaps or contradictions in their understanding.

Quick Knowledge Checks

In a large hall, you cannot rely on "nodding heads" to gauge comprehension. Leveraging polling tools (e.g., Poll Everywhere, iClicker) allows you to collect real-time data to decide whether to move forward or "pivot" your instruction. The data from a poll can help your next pedagogical move.

• High Mastery (> 70% Correct): Most students have grasped the concept. Briefly clarify the logic for the minority and move to the next topic to maintain momentum.
• High Confusion (< 30% Correct): The concept is likely too opaque the majority. Consider re-teaching the concept from a different angle before moving on and/or re-polling.
• Productive Disagreement Range (30%–70% Correct): According to Eric Mazur, this is the ideal state (“Golden Window”) for peer instruction. In this range, the concept is difficult enough to be non-obvious, but a significant portion of the class has the correct intuition and can "translate" the logic to their peers in novice-friendly terms.

Note: A common concern for faculty is that pausing for peer discussion will derail the syllabus. If you identify a "Golden Window" split but cannot spare the 5–7 minutes for a full discussion cycle, you can skip the peer talk and move straight to the expert explanation. For example you could say: “The room is split... Those of you choosing B are likely following [Logic X], which is a common trap because [Y]. However, if we look at [Variable Z], we see why C is the robust answer."

Building Community

In a large-scale learning environment, anonymity is one of the most significant barriers to student engagement. When students feel invisible in a course, they may be less likely to participate, ask questions, or interpret difficulty as a normal part of learning. Research on belonging uncertainty suggests that when students question whether they are expected to succeed, ordinary academic challenges can take on outsized meaning, with consequences for performance and persistence (Walton & Cohen, 2011.)

• Instructor Presence: Utilize the 10 minutes before the formal start of class to establish a social presence. Rather than focusing solely on technical setup, maybe play music, and walk the aisles. Small actions such as greeting students or informally engaging with early arrivals humanize the instructor and lower the psychological barrier for students to ask questions.
• Name Tents: Even in large rooms, using name tents or asking students to state their names before asking a question reduces the "chilly" climate of large halls.
• The "Warm-Call" Strategy: Cold-calling in a 300-seat room can raise the stakes of participation for many students, shutting down cognitive processing. Instead, use "Warm-Calling": "I’m going to give everyone 60 seconds to talk to their neighbor about this case study. Then, I’m going to ask a few groups to share what they found most controversial." This allows students to rehearse their answer, leading to more diverse and nuanced participation
• Shared Digital Workspaces: Use collaborative documents or discussion boards to allow students to co-construct knowledge in real-time.

Next Steps: Designing an Active Learning Lecture Course

The most engaged large lecture courses are ones that are designed to be active from the very start. If these suggestions have you feeling inspired to conceptualize a fuller course redesign, we encourage you to check out our Active Learning page. You’ll find strategies for even more robustly designing for student engagement from the very first day of class