This study explains a lot with regards to differential, and failure to differentiate exercise outcomes in people with low back pain.

Effects of abdominal stabilization maneuvers on the control of spine motion and stability against sudden trunk perturbations. (2007)

Abstract
Much discussion exists about which is the most effective technique to improve spine stability. The purpose of this study was to evaluate the effectiveness of abdominal bracing and abdominal hollowing maneuvers to control spine motion and stability against rapid perturbations. Eleven healthy males were posteriorly loaded in different experimental conditions: resting with no knowledge of the perturbation timing; performing each of the stabilization maneuvers at 10%, 15% and 20% of internal oblique maximum voluntary contraction with no knowledge of the perturbation timing; and naturally coactivating the trunk muscles when perturbation timing was known. An EMG biofeedback system was used to control the pattern and intensity of abdominal coactivation. The muscular preactivation of seven trunk muscles (bilaterally registered), the applied force, and the torso muscular and kinematic responses to loading were measured; and the spine stability and compression were modeled. The hollowing maneuver was not effective for reducing the kinematic response to sudden perturbation. On the contrary, the bracing maneuver fostered torso cocontraction, reduced lumbar displacement, and increased trunk stability, but at the cost of increasing spinal compression. When the timing of the perturbation was known, the participants were able to stabilize the trunk while imposing smaller spine compressive loads.

My comments:

One of the keys to good outcomes with low back pain patients is being able to sort through the research with regards to what is meaningful and what isn’t. When I think of spine stabilization I think of doing exercises in which my core musculature acts to keep my spine in a neutral position whether that exercise is a standing cable press or row, plank, squat, or deadlift. However in a number of (and probably most) research studies, spine or core stabilization exercise has come to mean isolation of the transverse abdominus muscle though abdominal hollowing, with some efforts to isolate the multifidus muscle for good measure. However when I was competing in Olympic weightlifting, none of us, EVER, made any such effort. We just lifted heavy keeping our chest out and our back straight (by straight we meant neutral) and lifted as hard as we could with no conscious effort regarding the contraction of any specific muscles.

What this study found was that efforts directed at abdominal hollowing had a near zero effect on the ability to stabilize the spine during an unexpected load, which was in contrast to simply bracing with all your muscles (as you would tighten them if you were about to be punched in the abdomen). As a former weightlifter, I thought Stuart McGill belabored his point a bit in his book. However as I review older and newer research regarding physical therapy treatments for low back pain, study after study still makes use of abdominal hollowing, calling it “spine stabilization” and “core stabilization,” and wonders why it does not work better than “general exercise.”  Often the “general exercise” groups are doing a great amount of genuine stabilization of the spine by holding the core rigid without any attention directed at isolating any particular muscle.