Disputas: Vigdis By Kampenes

Prøveforelesning

Bedømmelseskomité

Professor Barbara Kitchenham, Software Enigneering Research group, Keele University, UK
Professor Per Runeson, Department of Communication Systems, Lund University
Postdoc Margunn Aanestad, Department of Informatics, University of Oslo

Leder av disputas:  Knut Fægri

Veileder:  Professor Dag Sjøberg (UiO og Simula Research Laboratory), Tore Dybå (Sintef IKT)

Sammendrag

Norsk:
Empirisk forskning innen systemutvikling har i løpet av de siste 30 årene blitt en etablert disiplin. Den klassiske forskningsmetoden for å undersøke årsak-virkningsforhold er gjennomføring av eksperimenter. I avhandlingen evalueres aspekter av kvaliteten av eksperimenter som er publisert i løpet av en tiårsperiode og det gis forslag til forbedringer basert på funnene.

English:
Like any research discipline, software engineering research must be of a certain quality to be valuable. One way of assessing research quality is to conduct systematic reviews of the published research literature.

This thesis work is an assessment of the quality of published experiments in software engineering with respect to the validity of inference and the quality of reporting. The quality assessment was conducted by a systematic review of the 113 experiments published in nine major software engineering journals and three conference proceedings in the decade 1993-2002. The thesis provides suggestions for improvements, using the potential deficiencies detected as a basis.

The review revealed insufficient experimental quality, which has implications for inferences drawn from the experiments. This may in turn lead to the accumulation of erroneous information and misleading advice for systems development in industry. There is a need for an increased level of statistical power, increased use of effect size analysis, increased control for selection bias in quasi-experiments, and more complete and standardized reporting of these issues. More information about the experimental subjects, tasks, systems and settings is also needed. Implementing these improvements face certain difficulties, which are discussed in the thesis, as well as approaches for addressing them.

Kontaktperson

For mer informasjon, kontakt Lena Korsnes.