This report is submitted as partial fulﬁlment of the requirements for the Honours Programme of the School of Computer Science and Software Engineering, The University of Western Australia, 2005
Programming is a complicated skill to master, and learning to program is complex. The diﬃculty of ﬁrst year students studying computer science is that they generally do not already have a substantial understanding of computer programming. This results in either student retention, or student carefully selecting units in later years that have less exposure to programming. Associated with this is a desire to spot potential problem students have as early as possible. Timely warning can aid the teacher to provide more assistance to students. Furthermore, students would be able to change their current methods of study and make more informed choices before it is too late. The goal of the present research is to replicate previous investigations and test into whether the latency, or delay between certain keystrokes, correlates with the objective measure of programming performance. Controlled experiment was conducted in UWA with a total of 34 participants to test the hypotheses. Complete records of the keys being pressed and the millisecond timing were captured by using a logging tool known as User Action Recorder (UAR). These keystroke data were then recovered into digraphs according to their types. Spearman Rank Correlation Test was performed for each digraph type against programming score. This experiment had a goal to discover whether the previous typing pattern and results hold in a new setting. We examine the results from the experiment in UWA against the results from the previous two experiments. The results show that the correlation theory holds stronger in UWA. However, these results were not as signiﬁcantly strong compared to the previous studies. We further investigate this theory by analysing the additional typing data and found that correlation is only present while programming, and not while typing text. With additional development, these techniques may have promise for educational diagnostics.
Keywords: Digraph latencies, programming performance, keystroke latency, keystroke model, chunking CR Categories: H.5 ii
First and foremost, I would like to thank my parents for believing in me and stuck by me through tough times. Thank you for having faith in me. Secondly, I would like to thank my supervisors A/Prof Richard Thomas and Amela Karahasanovi´. Their guidance and assistance throughout the year have c been invaluable. I would also like to thank A/Prof Richard Thomas, the UWA Gleddon Visitor Program, and Simula Research Laboratory for making the experiment possible. I would like to thank Simula Research Laboratory for the materials, Kaja Kværn and Gøril Tømerberg for translating the experimental materials, and storing them in SESE, Gunnar Carelius for the technical support in Oslo. A big thanks to the UWA students who participated in the software engineering experiment in June, Ashley Chew and Laurie Mckeaig from the UWA IT support for conﬁguring the computers for the experiment, and Louise Bolitho for handling the payment to students. For all their support and advice, I would like to thank my fellow peers in the Computer Science department. I would like to extend my deepest gratitude to Robert Budiman, a dear friend who always supports me, and helps me ﬁnd my way on many occasions. Thank you Bob, I appreciate everything you have done for me. Finally, I would like to extend a special acknowledgement to Steve Draper in Glasgow for the Tinto’s paper.
Abstract Acknowledgements 1 Introduction 1.1 Chapter Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Literature Review 2.1 The Process of Transcription Typing . . . . . . . . . . . . ....