Cognitive Apprenticeship in Training for Conceptual Modeling

We take situated learning as the starting point for designing a training tool for an undergraduate course on conceptual modeling for information system design. The purpose of the tool is to create training scenarios that provide the same kind of problems and complexity which conceptual modelers face in the 'real world'. School training of conceptual modeling has mostly consisted of problems of well-structured character with ideal solutions. In this tool we want the students to practice in scenarios that resemble authentic environments while receiving some support in critical phases. The design of the tool is based on issues such as cognitive apprenticeship, authentic activity and language use, which have been in focus in the situated learning debate [Brown, Collins & Duguid 1989.], [Ramberg & Karlgren 1998]. Authentic activities as defined by Brown refer to "the ordinary practices of the culture" [Brown, Collins & Duguid 1989]. Our goal with this research is to evaluate how the cognitive apprenticeship learning model can be used to create complex and some what realistic exercise scenarios to facilitate training of the skills needed for conceptual modelers out in the 'real world'. We have used video recordings to study the problem solving activities of experienced modelers. The modelers were asked to create a conceptual model corresponding to a textual domain description, while thinking aloud. In order to ensure some degree of authenticity we picked a problem domain that was unfamiliar to the modelers, and also let them interact with a person acting as stakeholder and domain expert. The purpose of the video recordings was not to find out the general strategy used by experts to solve these kinds of problems. It was rather to gather examples of how experts could go about solving problems like this. A Scenario Environment Currently, we are building a prototype with two major components. The first component is an environment in which problem scenarios, such as financial risk handling or airline flight scheduling, are presented. In the scenario environment, the student can explore the problem domain by reading hyperlinked documents, posing questions to simulated domain experts, and watching video clips illustrating important and problematic issues in the problem domain. In the scenario environment there is also a workspace where the students create the actual model by drawing objects, relations, and attributes, etc. In the object-oriented community within system development the notion of patterns has been a buzz word the last couple of years, especially in object-oriented programming. Lately patterns have also become more popular in object-oriented design, by so called analysis or modeling patterns [Fowler 1997]. A goal of the course is to make students create general and reusable models and one way of doing this is by using patterns. Here we view analysis patterns as a language, which the students should practice to use, not as special constructs that they should memorize. Therefore we provide a pattern library from which the students can pick patterns to use in the workspace in their solutions. The patterns will also be accompanied by example ways of how they are used. By integrating the pattern library with the workspace we let the student use and modify patterns from the pattern library in the workspace and thus increasing the students' active use of the patterns. Apprenticeship and Scaffolding Features The second component includes some apprenticeship features and scaffolding support by providing 'tracks' illustrating how experts proceeded in solving the same problem, a pattern library, supported dialogs, and support for the workspace activities. One general feature of the workspace is that the possible ways of modifying, adding, and removing objects, attributes, relations, etc are restricted, i.e., only a pre-specified number of new objects can be created and the patterns can only be modified in certain ways. This makes it possible to keep track of where the students are in the modeling process and makes it more manageable to provide the apprenticeship and scaffolding features. We have used the video recordings in two different ways to ensure some degree of authenticity and to provide scaffolds for the students' problem solving. Firstly, to be able to show learners how experts go about solving similar problems. When learners run into a problem they can be shown video clips of how the experienced mo delers proceeded from the point where the student currently is. Secondly, we used the video recordings as a point of departure in the design of the system in several ways. One way was to create the master problem solving tracks, which students can follow if they feel they need guidance on how to solve the current problem. The tracks consist of 'steps', each corresponding to different actions taken, or decisions made, by the experts during problem solving. At each step the student gets to model the same parts as the expert did at this point. When the students are following a master track we can give feedback on solutions or partial solutions based on where the students are currently at in the track. We also use the recordings to extract critical points in the problem-solving process. We have seen that the experts all come to some common points of insight, e.g., that some generalization should be made or that some object should be split into several. We intend to use these insights to create 'stations' that the students should pass when going through a scenario. The purpose of these stations is to make it possible for students that do not follow any of the tracks to still get support in critical or problematic phases.