A before/after in a whole country is not as clean an experiment as the randomized trial done in this study. With the whole country, there could be a whole bunc

A before/after in a whole country is not as clean an experiment as the randomized trial done in this study. With the whole country, there could be a whole bunch of other variables changing in the before and after time frames. (Like maybe some new service comes online that results in people using a lot more bandwidth.) The design of this study allows the researcher to hold those things constant.