Unless you conduct a study where you divide peer groups blindly and send some early and send some later, it is just a (educated) guess this is the reason. Othe

Unless you conduct a study where you divide peer groups blindly and send some early and send some later, it is just a (educated) guess this is the reason. Otherwise, you can just walk trough data sets until you find one that meets your conclusion.