Search

The purpose of this document is to describe the design and execution of the quasi-experiment conducted in the Department of Computer Science, Virginia Tech, in accordance with the procedures described in the document "An Experimental Design for Evaluating SEES" prepared for NASA under contract NAS1-19610, Task 17. (Hereafter, that report is called the general design document.) The quasi-experiment serves three important purposes. First, the quasi-experiment is the proof of concept of the general experiment design. It shows that the procedures defined in the general experiment design can be implemented. Second, provides details for setting up a true experiment: identifying the research hypothesis, designing the investigation, selecting various variables, procedures, and controls, measuring the variables, and evaluating the results. Third, any insights revealed during the course of the quasi-experiment can be incorporated in the costly true experiment, thus providing a cost-effective experimental methodology. This report contains sufficient information so that the quasi-experiment can be replicated at an appropriate level of abstraction. We also document and interpret all the results of the quasi-experiment.

This paper presents the results of a study investigating the extent to which Independent Verification and Validation (IV&V) impacts early fault detection, and consequently, the mean time to remove faults. Two separate development groups, one with an IV&V team, are tasked to produce a software system from the same set of requirements. For each development phase, fault detection and removal data are recorded. An analysis of that data reveals that the group having the IV&V contingent: (a) detected errors earlier in the software development process, and (b) on the average, required substantially less time to remove those faults.

The purpose of this document is to describe the design and execution of the quasi-experiment conducted in the Department of Computer Science, Virginia Tech, in accordance with the procedures described in the document, "An Experimental Design for Evaluating SEES" prepared for NASA under contract NASI-19610, Task 17. (Hereafter that report is called the general design document.) The quasi-experiment serves three important purposes. First, the quasi-experiment is the proof of concept of the general experiment design. It shows that the procedures defined in the general experiment design can be implemented. Second, provides details for setting up a true experiment: identifying the research hypothesis, designing the investigation, selecting various variables, procedures, and controls, measuring the variables, and evaluating the results. Third, any insights revealed during the course of the quasi-experiment can be incorporated in the costly true experiment, thus providing a cost-effective experimental methodology. This report contains sufficient information so that the quasi-experiment can be replicated at an appropriate level of abstraction. We also document and interpret all the results of the quasi-experiment.