A study with students enrolled at an urban technical college compared learning of HTML coding across three Web-based program versions. One group received examples providing factual information about the source code (Facts group). A second group was presented with a step-by-step description of the source code (Procedures group). The third group was shown a visual model, including a diagram and arrows denoting the source code and the associated outcome (Visual Model group). A comparison of coding activity revealed that the Visual Model group constructed more correct code, with fewer trials, in the same amount of time. The Visual Model group also more accurately debugged flawed code. The Procedures group coded more correctly than the Facts group but differences decreased as problem complexity increased.