Abstract
Previous research on computer programming advocates that most computer science students, especially novices, lack programming competencies. The reasons given for this inadequacy is that most students lack the background knowledge, first experience of programming, and a new environment of writing programs in a syntax specific language, and so forth. Due to these reasons, the failure rate is high every year. Several researchers have used learning taxonomies; in that, Bloom's taxonomy has been widely used for assessment and learning of programming. Moreover, Bloom's taxonomy has been used as a scale for preparing the assessment questions, and the competency level was quantified based on that. In contrast, this study proposes a novel approach of programming assessment, in which the achieved competency level of a student is mapped to the respective cognitive levels of Bloom's taxonomy directly from the written code with no prior mapping of questions. The computation of the competency level in terms of mapping to the respective cognitive level is based on some principal criteria extricated from theories used in previous studies. Furthermore, this study emphasizes the basic topics of the structure programming course: Selection, repetition, and modular. The data collection was carried out from 213 students using an empirical test that is further analyzed through Structural Equation Modeling. The results show that Bloom's taxonomy is a beneficial tool for learning and assessing programming.