A lot of interesting experiments have been run with real implications for the future of development policy. But it's difficult as a graduate student to get involved with them. They take years to develop and are quite costly to run. Unless you're connected with some of the big-name faculty who have made their careers off of RCTs, they're impossible to do as a graduate student.

Experimental has been a good intermediary. The same toolkit, but a lot easier to do as a graduate student regarding red-tape and cost.