2207 Elimination of assignable cause of variation from quality performance data by applying "adjusted performance in postmortem"

Abstract

It is important to establish appropriate control limits based on past performance to conduct effective quantitative quality control in software development. We defined "adjusted performance in postmortem" to promote project members to identify an assignable cause of variation. "Adjusted performance in postmortem" is the result of eliminating the assignable cause of variation from the actual performance. By using "adjusted performance in postmortem", we can eliminate the effect of outliers to set control limits at the planning phase of the succeeding projects. We confirmed the elimination of assignable causes of variation by the application of adjusted performance in postmortem by statistical analysis of the actual data collected from our postmortem reports.