Analyzing the Decision Criteria of Software Developers Based on Prospect Theory
|Title||Analyzing the Decision Criteria of Software Developers Based on Prospect Theory|
|Publication Type||Conference Proceedings|
|Year of Publication||2016|
|Authors||Kina, Kanako, Tsunoda Masateru, Tamada Haruaki, and Igaki Hiroshi|
|Secondary Title||23rd IEEE International Conference on Software Analysis, Evolution, and Reengineering (SANER 2016) at Osaka|
To enhance the quality of software, many software development support tools and software development methodologies have been proposed. However, not all proposed tools and methodologies are widely used in software development. We assume that the evaluation of tools and methodologies by developers is different from the evaluation by researchers, and that this is one of the reasons why the tools and methodologies are not widely used. We analyzed the decision criteria of software developers as applied to the tools and methodologies, to clarify whether the difference exists or not. In behavioral economics, there are theories which assume people have biases, and they do not always act reasonably. In the experiment, we made a questionnaire based on behavioral economics, and collected answers from open source software developers. The results suggest that developers do not always act to maximize expected profit because of the certainty effect and ambiguity aversion. Therefore, we should reconsider the evaluation criteria of tools such as the f-measure or AUC, which mainly focus on the expected profit.
they sent a survey to developers and used flossmole to harvest the email addresses