SILC Showcase

Showcase May 2013: Exploration of CogSketch as an Instructional Tool in Middle School Science

Share this article using our url:

Exploration of CogSketch as an Instructional Tool in Middle School Science

Brian W. Miller, Jennifer G. Cromley, Nora S. Newcombe (PI)

Temple University

If you look at a refrigerator in any home having children, you will probably notice that children love to draw, and that they use drawing to help them process their thoughts and feelings about the world around them. Likewise, if you go to any science classroom or laboratory hallway, you will likely see that scientists also have a love of pictures and drawing. They too use drawing (and graphing) to understand and communicate their ideas and theories about the world.

Cognitive psychologists have suggested that drawing is a promising way for children to learn science (Ainsworth, Prain, & Tytler, 2011; Newcombe & Stieff, 2011). But there are some hurdles to using it in the classroom. To make sketching work as a learning activity, students need to be monitored regularly and given feedback as to whether they are succeeding (Corbett & Anderson, 2001). But imagine how difficult it is for teachers to give this feedback efficiently. Each child may draw the same object in very different ways. As the teacher walks around the room glancing at each picture to see how each student is performing, it is difficult for the teacher to determine what differences are trivial or surface differences and what differences reflect genuine confusion. Even if it took the teacher only one minute to analyze each picture, it would then take a teacher over half the class period just to check each student sketch once. That delay would leave some students lost most of the time. In comparison, writing is a constrained way of communicating. A few seconds is all it takes for a teacher to walk around the class and scan multiple-choice, fill in the blank, calculation or even short answer questions.

CogSketch is a software program that is capable of analyzing student drawings and offering feedback to students, guiding them to fix confusions and/or directing them to request specific help from the teacher or peers when they really need it (Forbus, Usher, Lovett, Lockwood, & Wetzel, 2011). This tool would free up the teacher to help struggling students while giving all students the kind of regular feedback that makes instruction effective.

CogSketch has already been used by undergraduate students in a geology class for learning about faults and about the carbon cycle (Yin, Forbus, Usher, Sageman, & Jee, 2010). However, CogSketch is a developing project with an evolving interface. One concern is whether the interface is easy enough to use and the feedback helpful enough to allow it to work with middle school students from average backgrounds. We set out to evaluate the current capabilities of CogSketch for use with such beginning science students.

We intended to use the program as a study aid, not to deliver original content. For this reason, we entered the classroom after students had completed their instruction on the target topic: the circulatory system. While they had received their usual instruction, the circulatory system is a topic with a great deal of detailed information and some very persistent misconceptions, including that it consists of one loop, instead of two loops with blood going from the heart to the lungs, to the body and back to the heart. We aimed to determine if CogSketch could help students further refine their concepts and strengthen their learning of vocabulary.

The intervention consisted of four worksheets designed to reinforce the names of the chambers of the heart, the flow of blood into and out of the heart, the movement of oxygen in and out of the blood, lungs, and tissues, and the double loop structure of the circulatory system. In each worksheet the students were given a simple set of instructions that included labeling and drawing arrows showing the movement of blood, oxygen and carbon dioxide (Figure 1). As they completed the worksheet they were free to ask for feedback from CogSketch whenever they wanted. When the worksheet was correctly drawn CogSketch gave them a special message, and they were allowed to continue to the next worksheet. They were given one hour to get as far as possible. About 70% of the students completed all four worksheets.

Figure 1
Figure 1

Before and after the intervention, students completed a three part assessment, which included labeling the chambers, the two-loop system, and the movement of oxygen respectively. The assessments used the same diagrams that were used in the CogSketch worksheets (Figure 2). There was a significant difference in the scores from pre to post testing for the first test (t[50] = 5.63, p .001, d = .789), and the second test (t[50] = 2.23, p = .03, d =.307), but not for the third test (t[50] = .837, p = .407, d =.115). Students also completed a survey about their experience. Students rated the activity highly (see table). In addition to Likert scale items, students also were free to give written comments. The most common complaint was that it was difficult to usethe finger pad as a drawing tool. This comment suggests that younger students may need drawing tablets to take full use of the CogSketch software. Other suggestions, such as improving the readability of feedback are already being implemented.

Figure 2
Figure 2

Perhaps the most important challenge for CogSketch in the near future is the development of teacher authoring tools. It is our hope that teachers will be able to design and use their own CogSketch worksheets so that this method of supporting student sketching can be integrated into existing curricula. Maybe one day soon, we can even find a way to put CogSketch on the refrigerator!


The measures used in this research were adapted from measures produced by Roger Azevedo supported by funds from the National Science Foundation (Award number DRL 0633918) and published in Azevedo, R., Cromley, J. G., Moos, D. C., Greene, J. A., & Winters, F. I. (2011). Adaptive content and process scaffolding: A key to facilitating students’ self-regulated learning with hypermedia. Psychological Testing and Assessment Modeling, 53, 106-140.