Cultural influences on requirements engineering activities

Requirements Engineering (RE) is a fundamental process in every software project and plays a significant role in ensuring software quality. It involves the critical activities required to accurately capture clients' requirements, completely and in line with users' needs. RE is a human-centric activity, and therefore, requires intense communication with software stakeholders. As culture plays a major role in the way individuals communicate and perform tasks, RE activities might be strongly influenced by individuals' cultures. However, the results of a Systematic Literature Review (SLR) confirmed that few studies had been conducted to explore the influence of culture on RE activities. Thus, cultural differences and influences need to be explored, as better understanding them might result in better RE quality and outcomes. The goal of this thesis is to explore the influence of culture on RE activities. We employed Hofstede's model and a mixed-methods design within two different cultural contexts. We adopted Hofstede's model because it is comprehensive and well-accepted model of culture. We adopted the mixed-methods design to systematically examine the issue by considering the strength of qualitative and quantitative approaches. The mixed-methods design comprises two phases. In Phase 1, we conducted 41 face-to-face interviews with practitioners, who were carefully selected to represent different types and sizes of companies, as well as several roles that are related to RE activities. In Phase 2, we conducted follow-up interviews with the same practitioners of Phase 1 to consolidate the collected data. Although the number of participants appears to be comparatively small, we are confident that we achieved adequate coverage (e.g., cover different sizes of companies, work experiences, roles related to RE activities and RE methodologies) and, therefore, data saturation. The interviews were conducted with practitioners from two cultures: Saudi Arabia and Australia. We selected Saudi Arabia and Australia by the following reasons: 1) the significant differences between them allow for a reasonable coverage of diverse values of cultural dimensions (as per Hofstede's model) which provides a sound basis for generalisation on an abstract level; 2) the cultural profiles of both cultures, as per Hofstede's model, have similarities to those of many other cultures (e.g., the United States, Canada, New Zealand, etc.), which make the results of this thesis could be applied to them as well. We identified 25 cultural factors that can influence RE activities. We investigated the implication of these cultural influence on RE activities, mapped them into Hofstede's cultural dimensions, and examined how these cultural influences might hinder or facilitate RE practices. We further investigated the practitioners' response based on three factors: the RE methodology adopted by the participants, the type of clients, and their company's size to provide a more in-depth analysis of the issue. Our investigation results were used as a foundation for developing a framework that maps between the cultural index ranges (as per Hofstede's model) and the identified cultural influences. Mapping the identifying cultural influence into Hofstede's dimensions allows us to generalise the application of our framework to identify cultural influences in any cultures that match the cultural ranges of Saudi Arabia and Australia. The purpose of the framework is to help RE practitioners identify and analyse the potential cultural influences they may encounter in local and global contexts. We introduced an cost-effective and practical framework to identify a set of potential cultural factors that influence RE activities in order to improve the process of RE activities. The idea was to develop a systematic way to identify cultural influences in different cultures, which is more simple and easy to apply than to conduct culture-related case studies. The framework also provides RE researchers with a theoretical foundation for more empirical research to examine the influence of culture on RE activities. To evaluate our framework, we employed three different approaches. We first conducted a comparative analysis with published case studies of other researchers. We intended to measure the precision and capability of the framework to predict cultural influences on other cultures that studied by other researchers, as identified in our SLR. The framework provided a higher level of accuracy (89% accuracy for the Thai culture and 75% accuracy for the Chinese culture). We then conducted another comparative analysis with Vietnamese practitioners to evaluate the framework with practitioners working in a cultural context different from those of our two main case studies. The Vietnamese practitioners determined almost all the highly expected cultural influences generated by our framework, as influencing RE activities in Vietnam. Lastly, we distributed a survey to measure to what extent the framework can help RE practitioners to identify and analyse cultural influences that impact RE activities. We evaluated the framework with RE practitioners because we wanted to develop a framework that practitioners can use without requiring any knowledge about Hofstede's model. The practitioners generally found the framework helpful in identifying and analysing cultural influences. The three evaluation approaches allowed us to determine the accuracy of our framework to identify and analyse cultural influences on RE activities. Moreover, the results assisted us to refine the framework. The framework provides a concise and systematic way to identify cultural influences in different cultures. The framework is simple, cost-effective and practical to identify a set of cultural factors that influence RE activities than conducting a culture-related case study. The framework will help RE practitioners communicate with software stakeholders more effectively and improve the implementation and quality of RE activities and their outcomes.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.