Moreover, we have found that the number of changes in reviews is independent of the reviewer in ConQAT. This unintuitive finding indicates that reviews, though people-driven, could be more people-independent than originally assumed. Instead, factors influencing how many changes need to be made inreview comprise the type and the volume of the task: Bug-fixing tasks tend to have fewer changes than tasks that implement new functionality. The more code churn or the higher the number of touched files in a task,the more changes do we expect during review on average.