Open Access

ARTICLE

A Risk Poker Based Testing Model for Scrum

Siti Noor Hasanah Ghazali¹, Siti Salwah Salim^1,*, Irum Inayat², Siti Hafizah Ab Hamid¹

1 Department of Software Engineering, Faculty of Computer Science and Information Technology, University of Malaya, Malaysia
2 Department of Computer Science, FAST-National University of Computer and Emerging Sciences, Pakistan

* Corresponding Author: E-mail:

Computer Systems Science and Engineering 2018, 33(3), 169-185. https://doi.org/10.32604/csse.2018.33.169

Abstract

In agile software development, project estimation often depends on group discussion and expert opinions. Literature claims that group discussion in risk analysis helps to identify some of the crucial issues that might affect development, testing, and implementation. However, risk prioritization often relies on individual expert judgment. Therefore, Risk Poker, a lightweight risk-based testing methodology in which risk analysis is performed through group discussion that outperforms the individual analyst’s estimation is introduced in agile methods. Keeping in view aforementioned benefits Risk Poker can offer, unfortunately, no study has been conducted to empirically prove its ability to improve the testing process to date. Therefore, this research is aimed at closing this research gap by (i) deploying Risk Poker technique as a risk-based strategy in the agile development lifecycle, and (ii) empirically evaluating improvement of the proposed test process. For this purpose, Risk Poker technique is coupled with test coverage for an innovated testing process in an agile project following Scrum in order to provide adequate test coverage for testing activity. A case study was conducted with 6 teams of undergraduate students to estimate test coverage using Risk Poker for an e-commerce system. Three teams estimated their user stories using Risk Poker, while the rest estimated individually and used an average to obtain the statistical combination. The results showed that the proposed usage of Risk Poker for risk analysis and estimate test coverage outperformed the averaged statistical estimation of risk analysis for user stories.

---

Keywords

---

Citations