For the most part, the first-time posters who submit one-word answers followed by a little filler to meet the minimum character count are not interested in anything more than participating in a quiz-show-format guessing game. They have no sense of what EL&U's standards are or what sort of resource the site is hoping to be—and admittedly single-word requests feed that impression by being (in most instances) fairly trivial questions that draw predictable answers and have little long-term value to the site. But single-word requests (and more recently, phrase requests) evidently have tremendous popular appeal on this site, drawing lots of page views and upvotes.
Some of the words suggested by newcomers without citation or documentation are legitimate suggestions, but many others are marginally suitable for the OP's context or are simply bad. The former, I think, deserve a chance to be improved, and we give posters a chance to do exactly that by adding a comment beneath the bare-bones answer recommending that the poster add a suitable definition from a reputable dictionary. Usually posters ignore the advice—but sometimes they respond positively, and it seems to me that those who do may very well become productive participants at EL&U, given their willingness to improve their initial answer, their respect for this site's standards, and their ability to provide basic support for their answer.
From a moderator's perspective, it must be a nightmare to have to slog through an endless marsh of crappy answers from clueless and sometimes belligerent one-off answerers. Nevertheless, I don't think that all insufficient answers should be treated the same. If a moderator feels that a word or phrase suggestion is unsuitable to the OP's context or is only marginally justifiable, AND if the suggestion appears without any documentation or other support, I think deletion is a mercy to everyone involved. Far too often, I've encountered a poor SWR suggestion in an Review queue answer and asked the poster to add a suitable definition from a dictionary to show why the posted answer may be helpful. This amounts to challenging the poster to find the Elephants' Graveyard or the Seven Cities of Cibola; and the only lesson the poster may learn from the experience is that, if you look a word up, you may find that it doesn't mean what you thought it meant.
On the other hand, some suggestions are good, and the only thing wrong with them is that they lack supporting references. In these cases, it seems to me, the problem isn't so much that the posted answer is irredeemable as that it shows annoyingly little effort or a real or feigned ignorance of EL&U guidelines for posting an acceptable answer. If EL&U were like some other Stack Exchange sites, we might get around the problem by having other site participants add the missing components to the answer. But that's not the way this site generally works.
Why? I can think of several possible reasons. One is that single-word requests are already on shaky ground here. As tchrist points out in his answer, very few SWRs have any staying power. Their contribution to the site is like an all-sugar lunch: they draw a big buzz for a short time, and then they crash. In the language of publishing, they have little or no tail. Why should anyone make the effort to improve someone else's SWR answer when the question itself is ephemeral?
A second possible reason is that, by any practical measure, an answerer who posts a quick-and-dirty SWR answer hasn't shown enough effort to deserve to reap multiple upvotes when someone else does the minimal work involved in meeting the very low standard required of a supported answer. The quiz-show aspect of SWRs, where in many cases there are a few widely known suitable answers that participants are sure to suggest sooner or later, already rewards quick-draw answers. If we were to adopt a policy of filling in the blanks on incomplete answers for those posters, we would surely encourage more of (or worse than) the same.
A third possible reason is that the subject matter at EL&U is not conducive to collaborative answer building. Questions about English usage or even about word or phrase origins are often better answered by separate responses that take different approaches to addressing the question than by a single canonical answer that addresses every aspect of the issue. But that being the case, we tend to consider answers (and answerers) as sinking or swimming on their own—both in SWR situations and with regard to more-complicated questions.
Conclusion
I think that if an SWR answer makes a legitimate but unsupported suggestion, we should give the poster a chance to amend it, with a warning that the window for doing so is narrow (specifying the time frame, if possible); then, if the poster doesn't improve the answer within, say, 24 hours, we should delete the answer.
SWRs are a gateway into EL&U for some people who eventually become major contributors to other types of questions. I wouldn't want to wall off that continuing source of valuable site participants. But if a one-minute SWR answer is off-base, I think we should be far more ruthless than we have been in the past in simply deleting it. In dealing with such an answer, advising the poster to find supporting evidence for the posted suggestion is a waste of everyone's time.
