Extreme Teen Reform
A lot of teachers and parents complain that some children are completely incorrigible and intractable to all efforts to change behavior, that no reward or punishment techniques work at all. "You just have to see these kids," they'll say, "Nothing I do could ever change them."
From their perspective, they're right. Of course, in reality, they're almost certainly wrong as an absolute proposition on behavior modification. If someone were literally unresponsive to all rewards and punishments they wouldn't likely survive very long. It would only be a matter of time before they stopped eating and drinking, put their hands into fire or a meat grinder. Not because they wanted to , but simply because nothing would stop them from doing so. All of these are "natural" reinforcers (for purposes of this post, I'll use reinforcers to refer to both reinforcers and punishers), but there's nothing special about them in quality over artificial reinforcers. Their advantages over artificial reinforcers are primarily in their consistency and the limitations of some dangers of resentment at punishment (though note that theists frequently invent a devil-figure to resent for natural punishers). This consistency tends to make them enormously powerful. Indeed, horror stories such as Steven King's Thinner work off of toying with these basic assumptions.
The real problem for the typical teacher or parent, then, isn't that the child is completely unresponsive, it's that they cannot identify or apply the reinforcer that results in the behavior desired.
The techniques and places where this can be done do exist, however. Places like Tranquility Bay, a "specialty boarding school" in Jamaica with a strict behavioral program. Looking at the overall program, to my eye it does appear designed to produce results. Apparently the kids are put on a stepped hierarchy, with each step indicating greater privileges. Those on step 1 are not allowed to stand up, sit down, move, or talk without permission. At step 3 they gain the right to a supervised phone call home. At higher levels they gain the right and responsibility to enforce rules on lower-levelled students.
There are other interventions as well. The "Observation Position" (or OP, as they term it) is lying, restrained, face-down until one admits their contrition. Once an hour the student gets a 10-minute break, and at night they sleep in the hall. An article describes this as potentially lasting months. Former students indicate it's "degrading, painful experience".
Clearly, even if one agrees with the necessity of the methods, opportunities are rife for mistakes and abuse. OP sounds as though it's not intended to be positive punishment (that is, it's not supposed to be painful except in the removal of all other privileges), but students' descriptions of it include being wrestled to the ground and having one's limbs twisted. I have a hard time believing the staff are trained sufficiently (turnover numbers would be telling) to be aware of the appropriate limits.
Humiliative punishment and tightly controlled rewards can be used in one setting: when one is trying to brainwash the subject. It involves breaking down all sense of self and self-worth and re-forming it, only allowing the subject to derive worth from the administrator. And that's exactly what these places are attempting. Parents who put their children in these expensive schools are putting them through brainwashing techniques worthy of cults and para-military establishments.
It still might be worth it, if a parent is convinced the child is headed on a course towards complete self-destruction. The Observer article suggests that for many of the kids, this isn't the case. Some of the drug and alcohol problems cited seem relatively minor, and one girl is noted as having been headed for Harvard, a straight-A high school graduate, before she made an "inappropriate choice of boyfriend." Family pressures keep her in even at 19, when she could legally leave without her parent's consent.
So yes, behavioral techniques can work, even for the hardened cases. A more difficult question to answer may be to what lengths we're willing to go to get them to work.