Use scripted conversations to fill knowledge gaps

When the same questions come up persistently during office hours, we can safely assume that our lectures are not addressing common knowledge gaps. This is a significant issue in mathematically dense courses, where gaps in foundational understanding can compound rapidly.

While blended learning and flipped classroom approaches have gained traction in higher education, much of the content delivery still relies on lengthy, monologue-style video lectures. These tend to replicate traditional instruction rather than transform it. To address knowledge gaps, we introduced a conversational video approach that involves recreating dialogues between instructor and learner. This has led to a more social, responsive and cognitively efficient learning experience, substantially improving comprehension, confidence and autonomy. 

Across six engineering groups, the intervention led to:

• A 55 per cent increase in mean test scores
• Large learning gains, particularly for lower-performing students
• Higher performance in both basic quizzes and more demanding exams
• 92 per cent of students recommending the method to peers
• High numbers of students reporting high levels of self-confidence (above 4.4/5.0 in surveys).

The groups who took part outperformed control groups in higher-order reasoning tasks, indicating deeper comprehension over superficial familiarity. 
 

How dialogue-based videos address learning gaps

1. They target misunderstandings through diagnostic-driven design

We began by identifying the concepts students were having difficulty with and transforming them into scripted dialogues where a “student” voice articulated the misunderstanding, and the instructor guided them toward correct reasoning. This structure helped students recognise themselves in the scenarios while modelling the cognitive steps needed to overcome confusion.

2. Multimedia elements reduce cognitive load 

We divided content into short, concept-focused units, using visual cues to reinforce key points. We kept the tone informal to ensure they resonated with students. This method addresses cognitive overload, the idea that working memory has limited space, and when we exceed it, it is harder for new material to be stored in long-term memory.

What distinguishes them from mini lectures is their emphasis on conversational problem-solving. This approach supported dual-channel processing (visual and auditory).

• Spotlight guide: How to motivate students and staff in higher education
• Resources on STEM teaching and research
• ‘GenAI and critical thinking can – and should – work together’

3. They provide flexible, self-paced learning opportunities

One of the greatest strengths of conversational videos is their asynchronous nature. Students can pause, replay and revisit content when they need to. This is an especially valuable feature for those who may require more time to process abstract reasoning. Survey results showed that 71 per cent of students watched the videos more than once. 

How to apply this approach:

• Identify the most common student questions
• Write short scripts as dialogues to address the questions
• Record short videos that focus on one concept at a time
• Use annotations
• Incorporate intentional thought-provoking questions
• Collect feedback after each video.

How it works in an engineering context

In one of our courses, students often struggled to interpret regression models. A common mistake was assuming that each coefficient represented a fixed value without understanding its real effect on the predicted outcome. In the video, the “student” asked what the coefficients meant. The instructor clarified the idea using a real dataset, showing with simple annotations how each coefficient reflects the change in the response while holding the other variables constant. 

This guided explanation helped students better understand multivariable regression and see how both numerical and categorical predictors influence model predictions.

Why this approach improves outcomes

The conversational format works because it combines three elements crucial for learning:

• Reduces cognitive overload through segmentation
• Enables learners to learn at their own pace, revisit content and integrate concepts more autonomously
• Connects theoretical ideas to real engineering problems, helping students see how theory works in practice.

Together, these components strengthen comprehension while narrowing achievement gaps, especially for students who may be hesitant to ask questions in class.

Scripted problem-solving conversations offer a powerful, scalable and student-centred way to reinforce learning in engineering and STEM courses. By directly addressing learning deficits, reducing cognitive load and enabling flexible study, the approach improves academic performance, boosting confidence and supporting autonomy. The results of this intervention demonstrate a clear path for institutions seeking to modernise instruction. Our advice is to start small, design intentionally and leverage this approach to connect with students in more meaningful ways.

Future developments such as shorter segmented videos, pause-and-solve prompts and live follow-up sessions will continue to refine this model and expand its impact across diverse learning environments. 

María Villarreal, Jobish Vallikavungal, Ana Sarmiento and Jonathan Montalvo are educators at Tecnológico de Monterrey, Mexico.

If you would like advice and insight from academics and university staff delivered direct to your inbox each week, sign up for the Campus newsletter.