The effectiveness of offshore software development: an exploratory study
Russell E. Blahetka, Judith L. Forbes · 2004
This dissertation explored elements identified in previous research, such as Metiu (2001) and Sheremata (2002), which dealt with various aspects of offshore software development. The purpose of this study was to perform a quantitative correlative study to determine the strength and nature of relationships between 15 variables and three posited factors representing management, schedule, and quality. The study also looked at the relative importance of these elements as described by managers involved with offshore software development and those managers not currently involved with the practice. The research posed two questions. First, how effective is offshore software development for software firms in terms of schedule and product quality? Second, what elements of offshore software development do managers consider the most important? Statistical analysis used Factor Analysis to determine the pattern of loading by the 15 variables in the study. The analysis attempted to use Confirmatory Factor Analysis and Path Analysis to confirm the posited model and determine causal relationships. To determine if managers agreed on the relative importance of the 15 elements hypothesized to be important to offshore development, the analysis included Friedman's Test. While the results indicated a potential three-factor solution, Confirmatory Factor Analysis could not confirm the three-factor model. Furthermore, the research failed to find a causal model that converged. The results of the Friedman test for managers involved with offshore development indicated general agreement as to the relative importance of each of the 15 elements studied. However, the study found managers not involved in offshore software development did not agree as to the elements' relative importance. The conclusions drawn from this study were that a three-factor model might indicate a potential model to explain offshore software development. Correlations for all four hypotheses were significant (p < .01). Three of the four correlations were in the hypothesized direction. The study concluded with suggested actions to improve the study, including changes to the instrument, increased sample sizes, and seeking additional variables to explain offshore software development.