What does it indicate when a statistical test finds a highly significant difference between groups, and the model used to test this difference explains all the variation in the data?