A Contribution to Conveying Quality Criteria in Mechanical CAD Models and Assemblies through Rubrics and Comprehensive Design Intent Quantification

Abstract

This research examined the use of assembly rubrics, described how they evolved from parts rubrics, and studied how they affect student self-evaluation. Instructor assessment of students was also evaluated, finding that while the assembly rubrics were partially understood and effectively used by the students, they were more successfully utilized by the instructors. Strategies designed to improve design intent communication in CAD models, in order to enhance their quality, with guidelines targeted to evaluate efficiency, have been addressed with this research. It is apparent that metrics directed toward the instruction of design intent are needed, since design intent transferred through CAD models can be performed at three stages with competing tradeoffs that must be balanced to arrive at the best modeling strategy. Research included the development of a validation approach that reflects that rubrics are valuable devices to expedite consistent design intent communication, and are vital not only for evaluation, but also for the communication of instructor expectations. This research examined how to clearly define qualities of design intent to enable easier CAD assembly assessment. It has been found that there is more inter-rater agreement and correlation between instructors than between instructors and students, for all rubric dimensions. There is strong to moderate correlation between instructors for the dimensions of validity, completeness, conciseness, and clarity, while slight correlation exists for the dimensions of consistency and design intent. Secondly, rubrics can also be described as being either static or dynamic. Static rubrics, existing in paper form only, do not provide immediate feedback to the learner. Dynamic rubrics perform calculations that provide immediate evaluative observations to the user. Besides, they can be independently adapted to specific situations depending on the capability of the user. Electronic rubrics are ideally suited for dynamic rubrics, and permit the use and development of both adaptable and adaptive rubrics, as described next. Thirdly, rubrics need to be adaptable which should make them easily understood and user-friendly, and adaptive (rubric can change itself, depending on the usage pattern). Evaluative rubrics are used when an expert determines the pedagogical progress of a learner, while formative rubrics are employed by the learners themselves, in order to chart their progress and identify scholastic deficiencies that are in need of remediation. Rubrics must be continually refined and improved, in an iterative, collaborative process, until satisfactory agreement is attained, both between raters, but also between raters and learners. Thus, assertions maps were developed, illustrating how the expand-contract strategy adapts the rubrics to CAD trainee progress, while assisting the understanding of the different rubric dimensions. Based on the assembly rubric experiments, it is apparent that the small differences between instructors suggests that the proposed assemblies rubric is sufficiently sophisticated to furnish an unbiased accumulative assessment of student performance. Accordingly, it can be confidently stated that raters can be used interchangeably without sacrificing accuracy. However, the assembly rubric possesses finite efficacy to produce formative self-evaluation of CAD assembly skills for new learners.