Prediction as a Teaching (and Learning) Tool

September 10, 2026

Welcome to fall 2026! I wish you all a great semester, wonderful students, and fulfilling teaching experiences. This semester, we celebrate 10 years of the Teaching Toolbox email newsletter. I am grateful for all those who have made the newsletter happen — Emily Gravett, who started the program and led it for nine years, the many contributors who took time out to write, the CFI leaders and staff who supported the planning and logistics, and, most importantly, you, the readers. Thank you for subscribing, reading, providing feedback, engaging in dialogue, and offering the occasional critical pushback.  

I wanted to start the academic year with a small and basic, but also impactful, focus. I got this idea from reading James Lang's book Small Teaching, which I highly recommend to you. As the title suggests, Lang focuses on little things we can do that make a real difference in our students' learning and that are based on solid learning theory. At this time of year, as our classes have started, our overall course design is set, and our mental load is at capacity, it may be a good idea to focus on little changes that we can still implement. 

One set of teaching interventions that Lang proposes uses students’ propensity to learn better if they can connect new knowledge to what they already know. In this context Lang refers to comparatively low-level learning outcomes such as remembering and understanding, but in ways that branch out into higher-level territory. For example, a simple pre-test that checks what students already know about the class topic primes their attention and can start a conversation about how new and existing knowledge relate to each other, thus stimulating analytic learning while also reconfirming foundational knowledge. 

I find it helpful that Lang centers several suggestions for connecting to pre-existing student knowledge on the idea of prediction: We give students a question or problem related to the course content and ask them to predict the answer to the question, or the solution to the problem, based on what they already know, followed by disclosure of the correct answer (or the range of correct answers). If done right, students will notice that they have already some knowledge that helps them make progress and that they are lacking important pieces of information, or central skills, that they still need to learn. Such "corrective re-modeling" (Lang, p. 25) of students' knowledge has several desirable results:  

• It gives students feedback that they need to learn what the class has to offer. 

• It provides them with information as to what it is they need to learn. 

• It can be used to introduce intriguing questions that motivate students to learn more (maybe even big, beautiful questions). 

• It can be used to spark students’ curiosity and increase their motivation. 

Here are some examples of how predictions can be used in class: 

• At the beginning of a class or class unit, have students complete a short test — open answers or multiple choice — that covers several key concepts or ideas that will be discussed. Students essentially predict the correct answers to the test questions. 

• As part of a reading assignment, ask students to read the subheadings in a text and formulate questions that they predict the text will answer. 

• In a large class, collect data from students about an issue (such as binge drinking in college) and ask them to predict how well their data align with nationally representative data, and why that is so. (Thanks to Robyn Kondrad for this suggestion.) 

• In a class that uses over-time data, show students a graph that ends at a point in time in the past and ask them to predict where the graph is going next. For example, in my U.S. Government class, I might show an old graph of presidential popularity ratings and then ask students to predict whether the line will go up or down, based on what they already know. 

• This can be extended to class subjects that do not necessarily have an over-time focus or are not necessarily quantitative. For example, provide students with a quote and ask them to guess the context. Have students predict policy opinions based on whether somebody is liberal or conservative. Or, in a science class, describe (or show) an experimental design and ask students to predict the outcome. 

• Show students a video clip, stop it, and ask them to describe what comes next. 

I am sure you can come up with more ideas based on your specific class questions, content, learning objectives, and disciplinary methods. (Chris Drew, at HelpfulProfessor.com, provides a list of further examples.) The implementation can be low-tech (take a piece of paper and write your answer) or high-tech (e.g. using PollEverywhere and other online systems); you may want to consult Robyn Kondrad’s toolbox on Using Polls To Enhance Learning. 

Whatever the set-up is, Lang emphasizes three elements: 

1. It is important to reveal the results, usually sooner rather than later. But it's also possible to use the prediction as a "cliff-hanger" at the end of the class to prepare for the next class or provide students with added incentives to do the readings. 

1. The reasoning behind the correct (or "correct") answer needs to be debriefed, to avoid reconfirming possibly faulty student beliefs and clarify any questions students have. 

1. The connections to the course material and the learning objectives have to be made clear. 

A prediction-based learning activity can be very simple (e.g. "write down three things that everybody should know about this topic") or quite involved (e.g. designing a hypothetical case study that includes a prediction activity). In either case, it is likely to improve student engagement — and learning — in your class. Try it out and let us know (for example email broschax@jmu.edu) how it goes!