Should computer scientists experiment more?

Abstract

Having recently participated in the Doctoral Consortium of the Australasian Computing Education Conference I was struck by the core methodological challenges faced by students. Initial proposals seemed to be along the lines of, "well I will need a control group and a comparison group to evaluate x or y condition or intervention." The corollary to this thinking then became "well how many students will be enough?" presumably so the findings could be generalisable in some manner. Yet do methods drawn from the natural sciences really furnish the most appropriate approaches for computing education research (CER), with its hybrid concerns of technology and people? If not, then which methods would be more suitable and how can students be guided towards their adoption? Development, 18, 2 (1999): 165-172.

Contributions of tester experience and a checklist guideline to the identification of categories and choices for software testing

“…1, the identification of categories and choices (or their equivalents) is also needed in domain testing (Beizer 1990), equivalence partitioning (Myers 2004), and in-parameter-order (Lei and Tai 1998;Tai and Lei 2002). Second, in line with the thoughts of the software community (Briand 2007;Carver et al 2008;Porter and Johnson 1997;Tichy 1998), observation of human performance is an essential element of software engineering. In this regard, our results should play a part in the contributions to software engineering research.…”

Guidelines for conducting and reporting case study research in software engineering

“…to compensate for measurement or modeling errors. Four different types of triangulation may be applied (Stake 1995):…”

“…social sciences (Robson 2002;Stake 1995;Yin 2003) which literature also has been used in software engineering. In the field of information systems (IS) research, the case study methodology is also much more mature than in software engineering.…”

Empirical Validation of MoDe4SLA; Approach for Managing Service Compositions

“…063.409 This paper validates our scientific solution (MoDe4SLA) for the real-life problem of managing complex service compositions through a controlled, quantitative experiment [3,4]. Experimental validation in Computer Science is recognized as being very important [5,4]. However, still a minority of research papers actually provides some experimental results [6,7].…”

“…However, still a minority of research papers actually provides some experimental results [6,7]. Without validating developed approaches like MoDe4SLA, however, a researcher might steer his research efforts into a fruitless direction [5]. In our evaluation we conduct an experiment with 34 participants.…”