On this page
Interactive simulations let you change a variable and watch what happens: drag a charge, tilt a ramp, add a resistor. One of the most studied collections is PhET, from the University of Colorado. Its research tells us both what simulations are good for and where they stop helping.
What did the PhET research find?
In a 2008 article in Science, Carl Wieman, Wendy Adams and Katherine Perkins summarised the project's findings. At that point PhET had "more than 80 interactive simulations", each tested with students.
The headline study was about circuits. In Finkelstein and colleagues' 2005 experiment, one group of introductory physics students did their DC circuit lab with real bulbs, meters and wires, and another used a simulation, the Circuit Construction Kit. The simulation group did better on a conceptual survey and also on assembling a real circuit and describing how it worked. Working with the simulation did not cost them hands-on skill on that task.
The Science article reports other results too. In a quantum mechanics course using a curriculum built around the Photoelectric Effect simulation, about 80% of students showed mastery of the concepts, compared with 20% in a course using traditional instruction. And from more than 250 think-aloud interviews, the authors drew a clear lesson: "students are not able to make sense of the science in the simulation just from watching. They must interact actively with the simulation."
Why do simulations help?
The PhET team points to a few design features:
They show the invisible. The article's example is the motion of air molecules in a sound wave; electrons and fields are shown the same way.
You control the variables directly. In the Wave Interference simulation, sliders set the frequency and amplitude of the wave, with an immediate visual response.
Several representations at once. The same sound wave as moving air molecules, as light and dark shading, and as a pressure-time graph.
They simplify on purpose. The article compares a good simulation to "training wheels on a bicycle". It hides irrelevant detail, like the colour of wire insulation that students in real labs have been seen worrying about.
Students also experiment more freely. In the circuits study, students using real equipment often stopped to ask the teaching assistant questions that showed concerns about hurting themselves or breaking the equipment. The simulation groups rarely did, and kept trying different circuit configurations to test their ideas.
What are the limits?
The same researchers are careful here. The Science article states plainly that "some learning goals are not addressed through the simulations, such as operating complex laboratory equipment."
The bigger limit is guidance. A 1998 review by de Jong and van Joolingen described problems learners can run into with discovery learning in simulations, and how instructional support can help. Richard Mayer's 2004 review in American Psychologist looked at three decades of research and found that, in each case, "guided discovery was more effective than pure discovery in helping students learn and transfer". A 2012 review of simulation studies by Rutten, van Joolingen and van der Veen found "robust evidence that computer simulations can enhance traditional instruction, especially as far as laboratory activities are concerned", but noted that most of that research ignored the teacher's support and how the simulation fits into the lesson.
So a simulation with no task is mostly play. A simulation with a question to answer is a lesson.
How to study with a simulation: a 20-minute routine
Take PhET's Projectile Motion simulation as an example. It lets you set angle, initial speed and mass, and add air resistance.
Predict (3 min). Write down: at what launch angle is the range largest? Does doubling the mass change the range? What happens with air resistance on?
Explore (5 min). Test each prediction. Change one variable at a time and note the result.
Explain (5 min). For each prediction you got wrong, write one sentence on why. For example, mass cancels in the equations of motion when there is no air resistance.
Connect to equations (4 min). Check one result against the formula. Range on level ground is u² sin 2θ / g, so without air resistance it peaks at 45° and does not depend on mass.
Transfer (3 min). Solve one textbook problem on the same idea without the simulation.
Step 5 matters. The aim is to carry the picture into problems you solve on paper, which is what exams test.
Where to find simulations
PhET's simulations are free on its website. The Fermi Simulation Lab has interactive simulations for the ideas that are easier to see than to read. For the physicist who founded PhET, read our profile of Carl Wieman and active learning. For when a plain diagram is the better choice, see simulations or static diagrams.
