RQ4: Is TestI a better predictor of conflict risk than TextI?
Tasks with non disjoint TextI interfaces more likely change a common file


TextI is a more precise predictor of conflict risk than TestI


Using acceptance tests to predict merge conflict risk