Creation of an online stoichiometry course that melds scenario based leaning with virtual labs and problem-solving tutors

Reflections on technology

In this section, we consider some of the processes and tools we used in creating the course since these had significant impact on the resulting product.

Above, we discussed the learning rationale behind our decision to use videos to set the scenarios and explain the concepts and problem solving procedures. Our past experience with creating video explanations had been with screen capture techniques. This was done by writing on a tablet PC (a pen enabled computer) and capturing the screen with Camtasia (For an example please see the Lecture M movies here). Given the difficulty of simultaneously writing neatly and speaking coherently, this process had evolved into one where the writing was done ahead of time. A small amount of writing was then copied onto a fresh workspace and erased in reverse order. The screen recorder was then started, with the instructor speaking while hitting the "undo" button as needed to cause the items to appear in forward order. This process typically requires multiple takes, since misspeaks and other errors are common. The process is quite time consuming; it takes between 10 and 20 hours to produce 1 hour of content. The level of professionalism in the is open for debate: it lacks the polish of typeset content but the handwriting may have the advantage of being more similar to a real classroom lecture and to what student work will look like when they carry out the computations themselves. However, we felt the biggest drawback of this technique is that the content is difficult to edit. Since the screen is captured as a bitmap, changing an element on the screen requires reproducing of all portions of the video for which that screen element is visible. Placing such a high barrier on changes would make it difficult to incrementally improve the course in response to reviews from learning scientists, educators and students, and especially changes indicated from results of learning assessments. Such iterative improvements are central to the mission of the OLI.

We therefore decided to use a multimedia editing environment to create the videos. To allow early review of the material by team members, we produced written scripts rather than capture impromptu verbal explanations. After review of the script, it was read aloud to create a sound track. The visual elements were then assembled. (The first movies were done in Final Cut Pro, which we later switched to Flash. The switch was motivated both by the smaller file size of flash movies and the fact that Flash is cross platform and familiar to more developers.) The style of the written scripts however deviated fairly strongly from the feel of an in person lecture. We therefore changed the process such that we began by filming an instructor teaching the material to a small audience. This was then transcribed and edited to produce a script which could be sent for review by team members. Although we did not do detailed time record keeping, the process probably takes 2-3 times longer than the screen capture technique discussed above. However, the content is considerably easier to edit, of which we took substantial advantage throughout the development and review process.

Overall, a middle ground between these two techniques would be highly desirable. It would be convenient to have a way to capture pen strokes and audio in a way that the content could later be edited in an environment such as Flash.

The tutor creation on the other hand, was not as tedious from a technology perspective. However, the technology does place substantial limitations on the resulting tutors. The short tutors were created with a set of Flash widgets developed by the OLI project team. These allow one to easily assemble a tutor by dragging components onto the Flash stage and setting parameters that specify the feedback and hints. For this course, a severe limitation is the lack of super- and subscripts in the hints and feedback messages, an issue we hope will be made easy to resolve by future versions of Flash. In addition, a more flexible set of tutor construction tools is available from the PSLC, called the Cognitive Tutor Authoring Tools (CTAT, http://ctat.pact.cs.cmu.edu/ ). A workshop on use of the CTAT authoring tools will occur this summer (http://www.learnlab.org/opportunities/summer/).

In both the OLI and CTAT tools, the user interface is a major challenge to creation of effective tutors. The interface consists of text boxes, drop-down boxes, radio buttons, etc. arranged on the screen. If one wishes to capture individual steps in the problem solving process, then one must arrange these screen elements in a way that supports entry of that particular step. This, in turn, gives students a large visual clue as to what to do next and so may not provide sufficient practice in how to generate the next problem solving step. Our current solution to this is the fading discussed above, whereby students are given a screen for entry of individual problem solving steps only if they fail to correctly solve the overall problem. A much better approach would be to develop a more fluid interface that allowed students to enter individual steps without providing them visual clues that over-scaffold their problem solving.