You're mixing up pullout torque (the maximum torque that can be produced before stalling, referenced in the multiple choice question) with electromechanical torque (the torque generated by the machine before rotational loss, the one you are showing the formula for).
So, as R_2 changes, the slip where you get maximum torque also changes (known as the pullout slip, s_po).
Substituting the pullout slip back into the electromechnical torque formula to find the maximum torque, all the "R_2/s_po" terms turn into "|R_th + j(X_Th + X_2)|", and suddenly the pullout torque is no longer dependent on R_2: the torque curve adjusts itself to create the same maximum torque at a different slip.