Defect content estimation for two reviewers

Claes Wohlin, Håkan Petersson, Martin Höst, Per Runeson

Research output: Chapter in Book/Report/Conference proceedingPaper in conference proceedingpeer-review

Abstract

Estimation of the defect content is important to enable quality control throughout the software development process. Capture-recapture methods and curve fitting methods have been suggested as tools to estimate the defect content after a review: The methods are highly reliant on the quality of the data. If the number of reviewers is fairly small, it becomes difficult or even impossible to get reliable estimates. This paper presents a comprehensive study of estimates based on two reviewers, using real data front reviews. Three experience-based defect content estimation methods are evaluated vs. methods that use data only from the current review. Some models are possible to distinguish front each other in terms of statistical significance. In order to gain an even better understanding, the best models are compared subjectively. It is concluded that the experience-based methods provide some good opportunities to estimate the defect content after a review.

Original languageEnglish
Title of host publicationProceedings 12th International Symposium on Software Reliability Engineering
Subtitle of host publicationISSRE 2001
PublisherIEEE - Institute of Electrical and Electronics Engineers Inc.
Pages340-345
Number of pages6
ISBN (Electronic)0-7695-1308-5
ISBN (Print)0-7695-1306-9, 0-7695-1307-7
DOIs
Publication statusPublished - 2001
Event12th International Symposium on Software Reliability Engineering (ISSRE 2001) - Hong Kong, China
Duration: 2001 Nov 272001 Nov 30

Publication series

NameProceedings of the International Symposium on Software Reliability Engineering, ISSRE
ISSN (Print)1071-9458

Conference

Conference12th International Symposium on Software Reliability Engineering (ISSRE 2001)
Country/TerritoryChina
CityHong Kong
Period2001/11/272001/11/30

Subject classification (UKÄ)

  • Software Engineering

Fingerprint

Dive into the research topics of 'Defect content estimation for two reviewers'. Together they form a unique fingerprint.

Cite this