I wrote a couple of days ago about the "broken windows" theory and education based on the recent article in Science. The theory essentially holds that small amounts of disorder will lead to more problems b/c people come to accept greater levels of disorder over time as it becomes the norm. And this certainly happens in schools as well as in neighborhoods.
But the authors make a bit of a finer distinction within the notion of disorder and norms. They divide social norms into "injunctive norms" and "descriptive norms" -- with the former representing, essentially, what people believe society deems acceptable and the latter representing solely what people observe, whether they believe it to be acceptable or not. The two are often in opposition -- for instance, if you think that speeding is a bad thing but also believe that almost everybody does it. In such a case I find it odd to call following speed limits a "norm" when you normally see people doing otherwise. I can be convinced otherwise, but I think the two would be better characterized as rules (whether written or unwritten) and norms. One dictates what is acceptable and unacceptable behavior and the other dictates what is normal and abnormal behavior.
The experiments test both how behavior changes when people witness rule-breaking and how it changes when people perceive disorder as a norm. In the case of the former, people are more likely to litter when they hear illegal fireworks being shot off than when they don't. And in the case of the latter, people are twice as likely to steal an envelope with money in it when the mailbox is covered with graffiti or surrounded by litter.
In other words, there are two different ways in which disorder can affect a school: In the case of rule-breaking, when students see other students chewing gum, carrying cell phones, or breaking other rules that might not really be enforced they come to believe that rules, in general, aren't really enforced or important and are more likely to break them. In the case of degradation of norms, when students see others running, shouting, pushing, etc. they're more likely to believe that chaos (or at least unruly behavior) is the norm and conform accordingly. Or at least that's how my experience would lead me to believe their research translates to schools.
So what are the implications? I'd guess the vast majority of teachers and administrators would agree, at least to some extent, with the previous paragraph. But knowing that disorder begets disorder and stopping disorder are two different things. I don't have all the answers, but I would say two things are most important:
1.) Don't make rules that can't or won't be enforced. If you're not actually going to suspend a kid every time you see them carrying a cell phone, then don't say you will. And in the case where the principal sets the rule this applies not just to you personally, but to the staff collectively. If all teachers aren't going to enforce the rule, then it's probably not a good rule to have. If kids see other kids playing on their cell phone in class, and those kids aren't suspended the next day then nobody's going to take that rule seriously. And once kids learn that not every rule has to be taken seriously, every other rule is in peril as well.
2.) Take action against low levels of disorder. That means somebody talking out of turn in class, yelling in the hallway, refusing to do assignments, etc. Maybe rules against such aren't codified, but when something small like this occurs it's important that the student is both made aware that it's unacceptable and that the behvaior is nipped in the bud to whatever degree possible.
Implementing number two is a task that I don't envy. What happens when a kid is made aware that his/her behavior is unacceptable and they continue it anyway? Some sort of consequences must follow. What happens when the student isn't bothered by the consequences and continues the behavior? Now the school's in a pickle (unless, of course, they can expel the student -- but even then such a decision should not be made lightly).