Regresar

Intelligent virtual environments to assist software testing learning: Meta-analysis for a family of experiments

Abstract:

The article reports the findings of the Meta-Analysis carried out to a family of experiments, in this analysis the effectiveness in detecting errors in the code is examined, with and without the support of a Collaborative Virtual Environment (CVE). The Family is made up of an original experiment (e1) and two replications (r1, r2) made at different times. As a synthesis technique, the individual and global effect sizes were estimated, using the weighted mean difference approach, as well as the determination of the degree of heterogeneity in the experiments of this family. The results suggest an equivalence in the effectiveness of defect detection with respect to using or not using an EVC, therefore it is suggested that the use of an EVC is an equally effective alternative to the traditional collaborative verification model (physical presence between members of a team).