Design Issues Adapting a Visual Paper-and-Pencil Test to the Computer: A Case Study--The Figure Classification Test.

This paper documents issues in converting the Figure Classification Test to the computer. The purpose of the test, which is almost entirely visual, is to determine the subject's ability to discover rules via the visual/spatial environment. The methodology of the paper-and-pencil Figure Classification Test is as follows: the subject views a series of two or three groups of pictures composed of printed shapes and is asked to classify each of the figures as belonging to one of three groups. In converting the paper-and-pencil test to the computer, the immediate concern was scanning the original drawings and converting them to line output. The philosophy of design for the project screens was to keep close to the paper version, yet to help the user navigate the test. The authoring aid used allowed mouse and keyboard input to operate equivalently for the user. Only minor changes to the original instructions were made; instructions to make the user familiar with the interface were added. Preliminary evaluation consisted of direct observation; 15 subjects at different times ran the program with no verbal instructions or coaching. The user interface and screen design seemed to be acceptable. No subject, after moving the mouse, ever attempted to use the keyboard for input. Users did not all gain immediate proficiency at the task, possibly due to the instructions. Device dependence issues also detracted from the computer test's equivalence to the original test. Direct observation of the program users appears to be a good first step toward improving and validating this computer-based test. (Contains 13 references.) (AEF)