This paper presents the findings of empirical research from 13 companies pa
rticipating in software development to identify the factors that may impact
software reliability. Thirty-two potential factors involved in every stage
of the software development process are defined. The study uses a survey i
nstrument to analyze these factors and identify factors that have significa
nt impact on software reliability. The survey focuses on the perspective of
the primary participants, managers, system engineers, programmers, testers
and other people involved in software research or development teams. Two t
echniques such as the relative weight method and analysis of variance techn
ique (ANOVA) have been used to analyze all factors and rank them in terms o
f their impact on software reliability. The research findings have importan
t implications for further research and the practice of software developmen
t. For researchers, it points to improvement schemes of existing reliabilit
y modeling and factors that may be further verified and extended in subsequ
ent research. For practitioners, it provides a general guide to the importa
nt aspects to consider in the whole software development process. (C) 2000
Elsevier Science Inc. All rights reserved.