I disagree. Perhaps in some areas, like electromagnetism or optics, but there are large fields in physics where it's not necessary (statistical mechanics, quantum mechanics, gravity, high-energy physics)
I am not saying that you should to do experiments in every area. Just that experiments in a few areas (usually mechanics/EM/optics/basic QM targeted by undergrad labs) is sufficient to give you the necessary intuition about the limitations of theory in all areas.
