Analysis comparing treatment-assigned and control-assigned groups regardless of whether treatment was actually received. The operationally relevant estimate for campaigns.