Any other cost estimation inhibitors?

Abstract

The purpose of this qualitative multiple case study is to research the current causes to cost estimation errors and revisit the research question of a well published quantitative study to explore whether responses would be consistent using qualitative methodology. Both overlaps and deviations are to be expected as a result of this comparison. The focus of this paper is on differences where new issues are likely to be found.The 8 resulting cross case cost estimation inhibitors validate 7 out of 16 issues found in the quantitative study. They are also related to the theory within and outside the area of software engineering resulting in five underlying software cost estimation inhibitors and providing a deeper understanding for the snapshot of the current causes of cost estimation errors reported by the practitioners.Lastly, the question of the practitioners' ability to report upon software cost estimation inhibitors is discussed. We argue that there might be causes of cost estimation errors that practitioners might be unwilling to reveal or might even be completely unaware of. Thus, they might unconsciously be providing researchers with invalid data.

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

“…The studies' participants perhaps assumed that lack of estimators' technical skills leads to inaccurate estimates, as it is more difficult to estimate software tasks you do not understand. This phenomenon is also indicated in further studies [Magazinović and Pernstål, 2008; Moløkken-Østvold and Jørgensen, 2005b;. But experience is not limited to technical skills.…”

“…The importance of using experience data and checklists also was explicitly researched and confirmed [Furulund and Moløkken-Østvold, 2007]. Additionally, not using a common template for estimates affects the accuracies of effort estimates [Magazinović and Pernstål, 2008]. Jørgensen's [2004c] analysis of different outcomes depending on the estimation strategy (top-down vs. bottom-up) also showed interesting results concerning the use of such analogies.…”

“…Poor requirements result in inaccurate estimates because imprecise descriptions lead to (possibly wrong) interpretations of the requirements, and thus, incorrect assessments by the estimator. The studies indicate that a weak or ambiguous requirement specification causes inaccurate estimates, whereas well-defined requirement specifications prevent projects from overruns [Furulund and Moløkken-Østvold, 2007;Grimstad et al, 2005;Jørgensen, 2004b;Moløkken-Østvold and Jørgensen, 2005a;van Genuchten, 1991;Magazinović and Pernstål, 2008]. This requirement is also reflected in a further study, as good knowledge on how to specify requirements can lead to more accurate estimates [Jørgensen and Moløkken-Østvold, 2004].…”

“…Lack of diligence by system analysts and programmers is mentioned as a reason for inaccurate estimation [Lederer and Prasad, 1995]. Experts in a second study mention inadequate time spent on estimation as a problem in estimation accuracy [Magazinović and Pernstål, 2008]. In cases of lack of diligence, the estimate accuracy may suffer as a cause of inaccurate requirement assessment by the analysts and a less valuable development by the programmers.…”

“…Jørgensen, 2004c 1 1 1 0.5 1 1 0.5 1 0 0.5 1 8.5 10 Jørgensen, 2010 1 1 1 0.5 1 1 1 1 0 0.5 1 9 11 1 0.5 0.5 0 1 1 0.5 1 0 0.5 1 7 12 Jørgensen and Grimstad, 2008 0.5 0.5 1 0 0.5 0.5 1 1 0 0 0 5 13 Jørgensen and Gruschke, 2009 1 1 1 1 1 1 0.5 1 0 0.5 1 9 14 Jørgensen and Halkjelsvik, 2010 1 1 1 0 1 0.5 1 1 0.5 0.5 1 8.5 15 Jørgensen and Moløkken, 2003 1 0.5 0.5 0 1 0.5 0.5 1 0 0 0 5 16 Jørgensen and Moløkken-Østvold, 2004 1 1 1 0.5 1 1 0.5 1 0 1 1 9 17 Jørgensen and Sjøberg, 2001 1 1 1 0.5 1 1 0.5 1 1 1 1 10 18 Jørgensen and Sjøberg, 2004 1 0.5 1 0.5 0.5 1 0 1 0 1 0 6.5 19 Lederer et al, 1990 1 1 1 0 1 0.5 0.5 1 0 0.5 0 6.5 20 Lederer and Prasad, 1995 1 1 1 0.5 1 1 1 1 0 0 0 7.5 21 Lederer and Prasad, 2000 1 0.5 0.5 0.5 1 1 0.5 1 0 0 0 6 22 Magazinović and Pernstål, 2008 1 1 1 0.5 1 0.5 0.5 1 0 1 0 7.5 23 McDonald, 2005 1 1 1 0.5 1 1 1 1 0 1 1 9.5 Table E1: Overview of the Quality Assessment Checks on the Primary Studies -Continued 24 Moløkken-Østvold and Furulund, 2007 0.5 0.5 1 0.5 1 1 0.5 1 0 1 0.5 7.5 25 Moløkken-Østvold and Jørgensen, 2005b 1 1 1 0.5 1 1 0.5 1 0 1 1 9 26 Moløkken-Østvold and Jørgensen, 2005a 1 1 0.5 1 1 1 0.5 1 0 1 1 9 27 Morgenshtern et al, 2007 1 1 0.5 0 1 1 0.5 1 0 0.5 0.5 7 28 Prechelt and Unger, 2000 1 1 1 0 0.5 1 0.5 1 0 1 1 8 29 Rombach et al, 2008 1 0.5 0.5 0.5 1 1 1 0.5 0 1 1 8 30 Subramanian and Breslawski, 1995 1 0.5 0.5 0.5 1 0.5 0 1 0 0 0 5 31 van Genuchten, 1991 0.5 0.5 0.5 0 1 1 0.5 0.5 0 0 0 4.5 32 Wesslén, 2000 0.5 1 1 0.5 0.5 1 1 1 0 1 1 8.5…”