A Systematic Mapping of Factors Affecting Accuracy of Software Development Effort Estimation

Abstract

Software projects often do not meet their scheduling and budgeting targets. Inaccurate estimates are often responsible for this mismatch. This study investigates extant research on factors that affect accuracy of software development effort estimation. The purpose is to synthesize existing knowledge, propose directions for future research, and improve estimation accuracy in practice. A systematic mapping study (a comprehensive review of existing research) is conducted to identify such factors and their impact on estimation accuracy. Thirty-two factors assigned to four categories (estimation process, estimator's characteristics, project to be estimated, and external context) are identified in a variety of research studies. Although the significant impact of several factors has been shown, results are limited by the lack of insight into the extent of these impacts. Our results imply a shift in research focus and design to gather more in-depth insights. Moreover, our results emphasize the need to argue for specific design decisions to enable a better understanding of possible influences of the study design on the credibility of the results. For software developers, our results provide a useful map to check the assumptions that undergird their estimates, to build comprehensive experience databases, and to adequately staff design projects.

Effort of EAI Projects: A Repertory Grid Investigation of Influencing Factors

“…Estimation experts themselves need to be considered as well. As prior research emphasizes, one of the most important aspects when it comes to reliable estimates in software projects is concerned with estimators’ characteristics (Basten & Sunyaev, ). Not surprisingly, the amount of experience of the estimator is important; however, also using a systematic estimation approach was said to increase the accuracy of the prediction.…”

“…Considering the high time and budget overruns in EAI projects (Rosa et al, ; Tuft, ), the relevance of project management is plausible. As budgets are typically calculated based on effort estimates (Basten & Sunyaev, ), a promising approach to avoiding EAI project failure is to scrutinize effort estimation in such projects (Rosa et al, ).…”

“…Effort estimation plays a crucial role in software projects in general (Basten & Sunyaev, ) and EAI projects in particular (Rosa et al, ). Effort estimates are used for project planning and staffing, success assessment, progress monitoring, and evaluation of developers (Basten & Mellis, ; Moløkken‐Østvold et al, ).…”

“…Despite decades of research concerning the development and improvement of formal models, empirical findings show that such models are seldom applied (Moløkken‐Østvold et al, ). One reason for practitioners’ preference of expert estimation over formal models is that the former approach can be more easily applied while providing at least an equivalent level of accuracy (Basten & Sunyaev, ). According to Jørgensen (, p. 55), “there is no substantial evidence supporting the superiority of model estimates over expert estimates.”…”

Antecedents and Consequences of Time Pressure in Scrum Projects: Insights from a Qualitative Study

“…These indices depend on each other (e.g., if the budget is reduced, this reduction has a direct influence on the quality of the software or the time). Related effort estimates are typically unrealistically low [16], and overruns of schedule and budget are common phenomena in software projects [17,18,19], which lead to time pressure.…”