Statistical analysis of an automated computer architecture learning environment

Abstract

The level of programming skill achieved by a method of learning introductory computer architecture with a strong emphasis on assembly language programming as part of a Graduate Diploma in Information Technology in the School of Computer Applications, Dublin City University is studied. The extent to which differences between students' backgrounds and study patterns, including usage of various software tools, influence the learning of assembly language programming is investigated. Information on the practical and academic backgrounds of students was obtained from an application form and questionnaire completed at the end of the course. Students' usage of the computer and software tools was monitored using log files and all this information was analysed to establish which factors impacted on the students' performance in programming examinations. Neither previous degree type nor degree class had a significant impact on levels of assembly language programming skill achieved but whether or not the student had experience of one high level language be.fore starting the course was significant. Surprisingly, it was found that those who did not prepare their programs in advance of going to the computer significantly outperformed those who did. The frequency with which students logged on to the system, the typical duration of their logon sessions and the frequency with which they used the XSPIM programming tool did not appear as significant, but the frequency of MIPSMARK test tool use and the order in which assignments were completed relative to other students were important.