There's a fundamental underlying cause to all of this: test-taking inherently measures the meta-skills that can be grouped as "test-taking ability". In a large part they measure the presence of certain attitudes and knowledge of the mechanics of the tests themselves. Students are not taught this, because the deconstructive attitude needed is seen as "disrespectful" of the authority of the curriculum's creator and implementer, and students are at the same time conditioned to believe that all information they are fed is always correct and equally valuable. This is not the case.
First, I want to explain what I mean by "disrespect" - I mean it as the view that any text you are handed in school is not derived from an absolute authority and crafted with the utmost skill. The text is there for you to question, for you to judge, for you to deconstruct, and for you to compare to past texts (that's why background knowledge is such a strong factor in test scores). Learning is a process of asking questions to develop an internal model of some complicated subject matter. We need this process, because our brains are smaller in capacity than the massive amounts of sensory information we gather from the world, so we develop effective forms of compression including "words".
A significant part of this is understanding the necessary degree of information fidelity that a particular class or standardized test will require - often times, pattern-matching a novel subject as a member of a superclass, then making educated guess from there. Or understanding what the "purpose" of a particular segment of the curriculum is, and using that knowledge to decide which information to prioritize remembering. And along a similar line, recognizing which material will be relevant to multiple questions and which details are so specific that they could really apply to a narrow set of questions that it is unlikely to see repeated on the same test. Most importantly, it is having the confidence to know when to cut a loss and focus on your strengths - you don't need to ever correctly answer 100% of the questions, especially on a standardized test where different grade demarcations are determined by clustering of the scores of the rest of the population. But you do need to answer all the questions, which is where all these heuristic methods shine and you can get some points by turning a 25% chance into a 50% chance, or reducing fidelity a few gradations and making the guess that corresponds to that superclass, or by having the vague recollection that this particular question asks about the piece of information who's purpose in the curriculum was to be an outlier and guessing accordingly.
And then the next step is to be able to decide when developing these skills will give you more value than trying to cram a few more details into your brain. But put less charitably, it means deciding when you know more than the school board officials and the less dedicated teachers, which they naturally do not like and do not want to encourage. If your goal is to pass the tests, it can be more efficient to study the generalizable skills of information processing versus class-specific information.
Ironically, the process of doing this will teach students better than most curriculums do, and from my experience it is those curriculums that lead students in this direction and tell them that it's okay to "disrespect" the system in this manner that have the best result. Not only are these meta-skills applicable to almost every subject, but they are critical in the real world where the truth of a statement is inherently contingent on its context or tautological. It is a shame that less dedicated teachers and school board members would see this as an attack on them, an attack on the validity of the system they have created - though I suppose in a way it is, haha.
However, attempts to teach this process usually fail because of the lack of the "disrespect" factor. "Find the main idea" is good advice - it is imperative to be able to quickly extract the purpose that a paragraph serves. But then students are taught to search for the topic sentence by highlighting words that are typically in "topic" sentences without explaining why, just that these words usually indicate that the sentence in question is a "topic" sentence. This is done for the purpose of answering the question on the standardized test "find the main idea". But now the students still aren't doing anything with the main idea, because telling students that they can use the main idea to construct a lightweight mental model of that paragraph, and then return to it to answer further questions that the recognize deal with the general idea of that paragraph.
Basically, schools are too afraid to tell students the truth that a) for schoolwork and tests you aren't supposed to "read" the material like you would read a book and b) not every single word in the material is equally valuable. And because of this, students are overwhelmed and never develop the fundamental logic and rhetorical skills that underlie most of the curriculum.
Of course, all of this applies because a student's worth is measured by a number, and therefore they are incentivized to either boost that number by any means or to give up because they put in a lot of effort and didn't get results, because no one told them they aren't actually supposed to put in that effort and that all the talk about hard work has an invisible qualifier that effort only matters when directed in a positive direction - that spinning your wheels gets you nowhere despite being even more tiring than just cruising along towards the goal. More penalties for not raising the number only works because it gives students more desperation to throw the notion of acquiring generalizable skills in the past to boost the number, which ironically lead them down the path of "disrespect".
This is what people mean when they say "the American education system destroys student's desire to learn" - it manages to turn the joy of discovery into a glorified exercise in sorting and pattern matching, where any effort put into exploring that doesn't lead to immediate quantitative results will always be outpaced by cynical deconstructions. We all love to say that "such and such famous inventor discovered one thousand ways to not accomplish their goals first!", yet also deride any instance of those "one thousand ways" in a number of different ways.