Paul Marshall shared an early version of his doctoral thesis with me that is important for science teachers teaching about moments, torque, and levers. His thesis compared physical and virtual environments for balance beam learning. The balances used are made from strips of wood with a fulcrum in the middle and weights that are hung from nails at fixed distances on either side of the fulcrum.
Marshall’s work is impressive and his writing style is refreshingly communicative without the vague obscurity of many contemporary researchers attempting to sound intellectual. Marshall presents key aspects of the long history, starting with Piaget, of balance beam studies. This is important for his coding scheme of the results. In this way, he goes beyond whether there was a difference in physical versus virtual representations of the balance beam on pretest-posttest content gains. I won’t attempt to describe an entire doctoral dissertation in a blog—but I will point out some salient findings.
In Marshall’s own words, “The main question to be addressed by this study was again whether using physical rather than virtual materials while completing the learning task would produce measurable differences in learning outcomes between the groups.” The subjects were 32 college students who worked in pairs on the physical or virtual tasks. They were given the pretest and then asked to explore the balance beam so they could perform better on the posttest. Students were limited to a maximum time of 30 minutes for the explorations. I consider this a “free exploration” situation rather than the typical “guided-inquiry” approach.
The results are similar to a study I described by Klahr, Triona, and Williams (2007). In both studies, all of the students had higher posttest scores, but there were no differences between the physical and virtual groups. Because all the interactions on the task were videotaped, Marshall was able to look at the type and duration of experiments. The physical group did an average of 28.8 experiments versus 37.2 for the control group, but these differences were not statistically significant. The mean time for physical experiments was 24 seconds versus 28 seconds for the virtual group, again not statistically significant.
Marshall classified the student experiments by how rich they were for providing useful data. There were no differences between the physical and virtual groups on this measure. Marshall also classified the experiments based on the Vary One Thing at A Time (VOTAT) protocol (Tschirgi, 1980). Again, there was no significant differences here either.
“In summary, there is little evidence that using physical materials in the balance beam task has a significant effect on participants’ choice of experiment search strategies. This contradicts predictions that subjects using the physical materials might engage in greater and more fluid collaboration with physical materials.” Once again, although for most teachers the virtual environment would be easier to implement in a classroom, the data suggests no significant differences between physical and virtual groups.
Marshall’s study went on to analyze the conversations between the pairs working on the physical and virtual tasks. There were no statistically significant differences in the amount of dialog that was coded as hypothesis, summary of data, prediction, alternate hypothesis, critique of a hypothesis, agreement with a hypothesis, extension of a hypothesis, justification by several experimental results, plan to test hypothesis, discussions about testability, arguments against a justification, or requests for explanation.
The virtual group, however, did have statistically significant more talk about suggestions of new experiments. This result was expected because when one student controls the mouse, there is more need for conversation about what they should do next.
The physical group had statistically significant more talk about description of results. Marshall suggests that this may have been because in the virtual situation both students could easily see what was happening, but in the physical situation, when both were sitting next to each other, each had slightly different perspectives.
“It must be concluded that using physical materials has little effect on learning about the concept of balance as presented in this study.” Marshall’s study adds to the nascent number of studies suggesting that virtual tasks make equivalent gains to physical tasks. Interesting wrinkles are the increased talk for planning experiments in the virtual environment and increased discussion of results in physical situations. Opportunities to extend all elements of discussions in all situations need to be explored.
Klahr, D., Triona, L.M., & Williams, C. (2007) “Hands on What? The Relative Effectiveness of Physical Versus Virtual Materials in an Engineering Design Project by Middle School Children,” Journal of Research in Science Teaching, 44(1), pp 183-203.