We are students of human-computer interaction who feel that speculation is important in the classroom. Speculating, failing, and then learning from our failure is how we innovate and advance as a society and grow as people. Science classrooms are important places for students to reason and speculate about the world, but sometimes they can feel intimidating to speak up in. We wanted to design a science demonstration focused on creating space for students to speculate about the outcome without fear of being wrong. Because we are not physicists or science teachers, rather than devising a new experiment, we decided to lift one out of some obscurity in the history of thermodynamics and show that through user research this experiment was still viable for classroom learning today, more than 200 years after it was first performed: Pictet's Experiment, which showed the reflection of radiative heat by mirrors and the apparent reflection of cold.
We chose this experiment because despite a simplicity that allows it to be shown in a middle-school classroom, it can be surprising even to working physicists and students with experience with physics and heat transfer. It can accompany a discussion of blackbody radiation, visible and invisible light on the electromagnetic spectrum, or merely a discussion of what makes things hot and cold.
To accompany this demonstration we provide instructional materials in different modalities, designed through iterative user and concept testing of students' inclinations to speculate and reason about the outcomes. These are integrated into a short lesson plan with suggestions for staging and key questions to ask. We also offer guidance on how to implement the lesson, materials to support student engagement, and evidence for the impact of a learner seeing the demonstration for the first time.
This project contains a capsule lesson plan appropriate for middle and high school science classrooms with a heat transfer component. The plan includes directions for constructing and conducting a demonstration, key concepts and questions for students, as well as a variety of optional discussion questions to allow the lesson to be tailored for student level and time allotted.
This lesson plan is released under a public domain license. Anyone may freely copy, modify or re-use the material we have created.
If you want to demonstrate this experiment in your classroom you may follow the links below.
- Lesson Plan and Teacher's Guide
- Experimental and Material Setup Instructions
- Resources for replication
Read on for more information about why and how we put the lesson plan together.
We did not want to make an app or design a change to how classrooms work. Science teachers don't need us or technology to help them understand that speculation is important. Classic experiments with pendulums are widely used to help students reason and speculate about concepts like mass, and science educators have been writing for some time about the importance of demonstration in the classroom.
We wanted to take advantage of our background as designers and researchers as well as the opportunity we had to conduct user and concept testing with students in order to attack a problem of narrow scope: a lesson plan that helps prompt speculation and inquiry in students, where the key stakeholders are limited to instructors themselves. This allowed us to focus on stakeholder needs and produce a solution that could be deployed today just as we have designed it.
Although Pictet's experiment was regularly demonstrated to students and included in at least one popular teacher's guide in the 19th century, someone today is likely to find it (if they find it at all) recorded as a curious footnote in the history of physics. Sandwiched between discussions of "frigorific rays" and "phlogiston", its value as a teaching tool is obscured. Further, finding pragmatic information about the demonstration such as suitable materials or valuable practices is a challenge and requires access to and a willingness to read paywalled journal articles for an uncertain benefit.
We suspected that middle and high school teachers were unlikely to see this as a good use of their time and we aimed much of our early secondary research at lifting Pictet's experiment out of this historical frame and collecting information about its use as a teaching tool into one place. This notion turned into our idea for our final deliverable for this project: a lesson plan. With our lesson plan, an instructor who discovers the experiment in one of those footnotes mentioned above and searches for more information will have, readily available, enough information to determine both the value of demonstrating it to their classroom and practical information about how to do so.
Teachers often have very few chances to test out additions to the classroom, but we could test and iterate on this specific goal with users throughout the design process. This allowed us to bring to bear our experience with user testing and validation to the problem of designing a lesson plan.
Because our solution is aimed at middle and high school teachers, the best user population is students of those ages. However, minors are always considered a vulnerable population for research (including class projects!) so all work must proceed through an Institutional Review Board (IRB). Given that the time frame for our project was less than 10 weeks, we moved forward by testing important elements of the lesson on a proxy–university students. On top of this limitation was a further one, we could not test users in an environment close to what we expected the lesson plan to be used in: a classroom. Our user testing was built around the recruitment of individuals for interviews and demonstrations.
Given these limitations, we approached user testing with the aim to first validate that our demonstration could create a space for speculation by the user and then to observe how user responses changed as the demonstration evolved. As the project went on and we continued to iterate on our prototypes, we put much of what is the lesson plan in front of users for validation and concept testing. However, the whole of the lesson plan could not be tested in the classroom environment we expect to see it used in.
Finally, we are not students of physics or working physicists. We feel experimental demonstrations of scientific knowledge must be carried out by scientists and non-scientists alike in order to broaden the scope and character of scientific knowledge. Our work in building the experiment and attending carefully to the user's behavior can offer lessons for science educators, not least of which is drawing attention to an experiment we believe has been unfairly neglected.
We not only needed to read up on the experiment, but conduct user research as well. For example, our secondary research showed that Pictet's experiment was able to puzzle and surprise even well-educated physicists at the time, but we did not know if students today would react the same way. Would students who live in the modern world of heat lamps and infrared cameras still have their intuitions challenged? Did this experiment in fact create a space where students are able to speculate freely?
To answer these questions and others, we conducted user interviews with key stakeholders and moderated user testing with various iterations of prototypes, both of which we will explain in detail below.
Our stakeholders for this project–the people who we hope will take up this lesson plan–are middle and high school STEM teachers. We conducted semi-structured interviews with three teachers via Zoom. One teaches in a well-resourced private school, one in a home-school co-op, and one via community science education programs. They were asked various questions regarding their perspective on demonstrating science experiments in front of students and general thoughts about the best way to create a space where students feel comfortable to speculate openly without the fear of being wrong or judged by their peers or teachers.
User interviews with undergraduate and graduate students were conducted as semi-structured interviews at the student union building at a local university. They consisted of questions asking about general perspectives and experiences about science classes and also specific questions regarding how different situations impacted their ability to speculate freely within the classroom. While we could not interview middle or high school students for our work, speaking with a mix of students inside and outside of STEM fields helped us understand how to consider students as a user group in general, rather than just speaking with instructors.
Once we conducted initial interviews with students, we prototyped virtual and in-person demonstrations with university students at the student union. Each demonstration included a short pre- and post- interview which we used to evaluate the effectiveness of the staging and presentation of the lesson.
Through the synthesis and analysis of our qualitative data collected from the user and stakeholder interviews, the common themes and trends that emerged helped guide major design choices. For instance, our initial goal was to replicate the experiment and produce only a video to be shown in the classroom. While some of our stakeholders wanted options of how to demonstrate experiments to the class including simply showing a video, students prefer live demonstrations or hands-on experiments as opposed to watching a video. One student noted a common sentiment, saying "if this is something that can be demonstrated in person it would definitely be more interesting." This along with other discussions with students who preferred seeing and interacting with a demonstration even if they were not responsible for performing it, led us to produce a lesson plan with instructions on guiding the demonstration.
Another way that we involved the input of the interviews was with the worksheet handout that we included in the lesson plan. One teacher explained their philosophy of trying "to get students to make a prediction first as a way to engage more deeply with what's about to happen, especially if it's a common misconception that we are trying to reveal." This sentiment was echoed by a student who shared how they are: "[more comfortable with] 'write down what you think would happen' or make it anonymous [type of situations]." With this in mind, we crafted a hypothesis worksheet that had the students anonymously write down their hypothesis of what they thought was going to happen in the demonstration and why. Then after the demonstration, on the back of the same worksheet, it asks the students to explain what actually happened and what if anything was surprising to them.
When we started on this project, lots of things were uncertain to us. We did not know whether or not students today would find the experiment counterintuitive or if it would provide space for them to speculate about the phenomena at work. We also did not know if we could recreate the experiment ourselves; how to integrate our secondary research about inquiry-based learning with our user research; or whether or not low and medium-fidelity prototypes would give good approximations to what we wanted to test.
Our first major takeaway from the project was that low-fidelity prototypes can work well in this space. Users, especially those we engaged in person, had less difficulty than we imagined understanding papercraft representations of the experiment, allowing us to effectively perform concept and user testing early on in the design process. Virtual demonstrations proved less effective in terms of eliciting speculation but still provided important feedback which we incorporated into our design.
Another takeaway came from the nature of the user tests we were able to perform. We discovered in our user and stakeholder interviews that students preferred to give speculations and hypotheses first in private or in small groups before volunteering these ideas to the whole class. Motivated by this, we developed worksheets to go along with the lesson plan where students can record speculation privately. However, our user testing was almost exclusively one-on-one, and users were understandably reluctant to write detailed feedback during user tests–preferring to offer it verbally. This meant that while we have good evidence from user and stakeholder interviews on the value of private, written speculation in these scenarios, we were limited in how we could assess our solution here.
The limitations noted above offer clear signs for future work. Our next steps for the lesson plan are:
- Seek out Institutional Review Board (IRB) approval to demonstrate and assess the response of our intended user group (middle or high school students) in a classroom environment.
- In collaboration with stakeholders, develop assessments for student learning during the demonstration. As one of our stakeholders noted, "Most of the learning is answering post-lab questions."
Aside from these major improvements, our lesson plan is available here on GitHub, where comments and suggested changes can be continuously integrated. The lesson plan and associated materials are shared under a free license so anyone can copy, adapt, and re-use what we have created for their classrooms.
Contributions are welcome!
- This project is the work of Adam Hyland and Daniel Teramoto, students at the Human Centered Design and Engineering department of the University of Washington.
This project was produced in Dr. Daniela Kim's User Centered Design course. We are grateful for her feedback and support as well as the support of our awesome TA Brian Kinnee. Thanks also to Dr. Brock Craft for assistance with fabrication.