"It's perfectly valid criticism [to point out] an obvious flaw in [someone's] writing. For instance, if an author writing on the role of race in welfare who completely ignores anybody who's not white or black."
No, that is untrue. You can write about race in welfare and only consider two races. Food for thought:
P1: No single work of any kind will ever be utterly comprehensive;
P2: This means that, in detail, there will always be something more that can be said about a subject;
P3: Whenever an academic produces a work, it will therefore not be comprehensive;
P4: Whenever an academic produces a work, there will always be something more that can be said about it.
So, either all academic works are failures before they begin, OR, it is possible that you can examine something and contribute usefully to a dialogue without saying everything that can ever possibly be said.
Ergo, anytime any work (academic or otherwise) comes out, it is a non-sequitur in the worst sense of the term to say "The author did a poor job because he did not also discuss _________." Only if the author claimed to be producing a completely-comprehensive work could he be criticized so. There are only so many pages, and so much time, in the world.
You can discuss welfare race roles using white and black alone. You can also discuss them using Inuit and Chechnyan. Both studies, if done well, could contribute knowledge to the overall picture.
Imagine that you did a welfare race study that took into account, say, 100 races. And then, when asked about your work, someone called you a poor writer because you did not take into account the 101st race?
How far does "comprehensive" have to go?
Answer: it can never be attained, and academics are going to keep bitching in literature reviews* about whichever nitpicky part of something wasn't discussed simply because if they stop producing articles to that effect, they won't get their tenure. A few brave souls have tried to challenge the system, and emphasize teaching quality and original research over that kind of crap, but they were handily shot down by the book and journal mills, and the future does not look bright.
* (Literature reviews defined in brief: before you are allowed to write about anything, prove that you know what you are talking about by describing the last X books written on the subject by people who had accepted positions within any given debate by virtue of their professorial jobs at the right universities at the right times. Most academic articles, and especially books, are considered incomplete without a sizable literature review to prove that they are worth reading. Nach, academics spend a lot of time writing literature reviews about the same literature in the same terms, over and over and over...)
That kind of discussion--bloviating about what was not said--distracts from what WAS said, and discourages floating original ideas. Original ideas do not necessarily require the support of seven pre-existing professors at seven prestigious universities.
The absolute terror to some academics is the thought that someone, somewhere, might have a great idea without having been molded through the western university system first.
Novel, dynamic ones would shatter the old, or reconceptualize it in such a way that it would be crippling to try to link them. And if you waste time tying original ideas back to 10 other respected articles in the field, research space is lost in favor of yet another literature review. Hasn't Harvard gotten enough head, yet?
Life is dynamic change. Academic ritual tries to hobble that change to the tenure publishing system, so that there is a new crop of books to churn out of the press and sell at $60 apiece to each new class of undergrads with predictable regularity.
It is an awful habit, when presented with anything nonfiction, to postulate in learned fashion about all the things it did not cover. There will always be something not covered. It is banal to spend time talking about that, even though it passes for learnedness among western academia--every new idea you hear is not an opportunity for thought, but instead an opportunity to recite what you have most recently read or are working on. It is reassuring, when you see someone else publish something new to you, to be able to critique it for not having cited to the last three articles you read on the subject. It's the same defensive habit as snickering at someone who walks into the room wearing a striking new fashion choice: "Oooh, she thinks she's so unique, but what a wasted attempt--she failed to take into account Prada and Gucci." Repeat ad nauseam with music, social decisions, marketing departments, or articles about the early twentieth century Russian Revolution.
Academics cling to this kind of behavior because it justifies their jobs, and the system in which they operate. If they weren't busy pointing out that no single work is comprehensive, then, why, normal people might be able to become authorities on things! People who haven't gone through the western university mill! Onoes the end of life on Earth!
Take it from Karl Mannheim: people (even academics, despite their massive educations!) tend to believe ideologies that support their economic well-being. For an overly lengthy and complicated discussion of this, see Mannheim's Ideology and Utopia.