1. What was the artist attempting to do?
2. Were they successful?
3. Was it worth doing?
That there would be analogies between art evaluation and research evaluation is virtually inevitable; they are both concerned with skills of production, and thus naturally organized in means-end terms. This is precisely what is going on with the three questions: (1) is a determination of the specific ends of the work; (2) is an assessment of the fitting of means to those ends; and (3) is an assessment of how those ends relate to more general ends. Any account of how any practice is evaluated will be at least broadly analogous to the evaluation of art. So one would expect there to be such similarities to works of art if we focus on 'works of research' (which we arguably should).
But Andrew Gelman at the link suggests a possible problem with this:
There are many cases of successful art, and for that matter successful research, that were created by accident, where the artist or researcher was just mucking around, or maybe just trying to do something to pay the bills, and something great came out of it.
I’m not saying you’ll get much from completely random mucking around of the monkeys-at-a-typewriter variety. And in general I do believe in setting goals and working toward them. But artistic and research success often does seem to come in part by accident, or as a byproduct of some other goals.
One old phrase for this kind of situation, in which something progresses by fulfilling ends at which it did not aim, is 'heterogony of ends' (Heterogonie der Zwecke), due to the German moral philosopher and psychologist Wundt. As he puts it:
The resultants arising from united psychical processes include contents that were not present in the components, and these new contents may in turn enter into relation with the old components thus changing again the relations between these old components and consequently the new resultants that arise from them.
Thus, for instance, in the course of acting according to one set of motives, and by the very process of acting according to them, we develop another set of motives, which do not replace, but interact with, the motives we had before. The development of new goals is a natural side effect of pursuing goals.
There is, on the other hand, some reason to be cautious about the introduction of accident into the mix. Art (like research) is not something that happens unawares. If a dog shakes off water, the waterdrops may make a pleasing pattern; but the pattern is not art if nobody notices it. So the mere accident doesn't accomplish anything. But a photographer might present it artistically, or a painter might represent it artistically, and then we have art. Art (and research) involves skills, and it is the application of those skills that actually makes anything art (or research). Thus, one might say, the accidents one might name as contributing to art are in fact just vividly dramatic examples of what artists really are always doing -- take the contingent features of their material and the situation in which they find themselves, and use them artistically. But it is in fact the skill that makes the art, and skills are analyzed by means and ends.
We also have to keep in mind that accidents can be only partial and ends can be complex. Our ends are never simple monads, but have their own structure. A bare end (world peace) is nothing but a wish or velleity, if even that, but when we are actually doing something, we have hierarchies of ends. Even analyzing something as simple as deliberately tying one's shoes before a race turns up an entire structure: I make this loop to make this knot to keep my shoes cinched to keep them on to avoid tripping to run better to compete in the race, etc. Thus an accident may well be unexpected and even inconsistent with one end, but allowed for and consistent with a more general end -- for instance, I may be looking for Such-and-such Street to get to the park, but in the course of doing so come upon an easier route. Did I find the park by accident? Well, in a sense yes and in a sense no: I was looking for the park; I was trying to do it one way but discovered a better way; the better way came by accident, but would not have been discovered at all unless I had already been looking for the park.
Thus we have a sort of aporia here -- on one side heterogony of ends and on the other analysis of skill, each of which can be taken as suggesting an opposed view of the role of accident. But I think we can deal with it fairly easily, by asking the question: In terms of what is art (or research) successful? To be successful is to achieve ends or goals. In the case of the researcher mucking around just "trying to do something to pay the bills" it is nonetheless not a matter of sheer chance that the researcher was examining this rather than that, or that, having found it, they made this use of it rather than that. Even with the researcher just "mucking around", it's not chance that they were mucking around with this rather than that, or doing it in this way rather than that. There's already a rather robust structure of ends in place -- standing operational goals, so to speak. And non-chance this-rather-than-that means that we are still comfortably within the realm of evaluate by means and ends.
And what questions can we ask to evaluate by means and ends? Nothing other than: What were they? Were the means good? Were the ends good? And those are noticeably just more general forms of the three questions above.
ADDED LATER: I intended also to note, but forgot when actually writing the post, that often the contribution of accident to art (or research) can easily be accommodated by the third question -- sometimes an accident makes a line of research more worthwhile than one could have ever expected, because it put you in the right place at the right time with the right resources at hand.