One of the things I notice with my students who have trouble with assumption questions is that even though they can often articulate the difference between the question types, when it comes to attacking the questions, it’s clear that they’re conflating the significance of the two types of assumptions. It’s very similar to a common mistake made on logic games, so let me address that first, by way of analogy.
Sometimes on a “Must Be True” question, (in the games section), you have to test answer choices – try to build a layout. So I’ll see a student test, for instance in a sequencing game, answer like, “Bob finishes last” by putting Bob last, just as they would on a Could Be True question. This is completely wrong, because even if it works (a valid layout can be created), that only shows that Bob COULD be last; it gets you no further toward figuring out whether Bob MUST be last. Instead, put Bob somewhere OTHER than last; if *THAT* works, then you know it’s the wrong answer; on a Must Be True question, you eliminate answers by finding counterexamples.
OK, back to LR and assumption questions. Let’s say an answer choice on a necessary assumption question is, “All dogs are dangerous.” When I ask them to articulate why they chose or eliminated that answer, they’ll say, “Well, because if all dogs are dangerous…”
And the problem is revealed. Because on necessary assumption questions, it’s completely irrelevant what happens if the answer choice is true. Necessary assumption questions ONLY care about what happens if the assumption is false (hence the negation technique). Specifically, which answer kills the argument (if false)
Conversely, sufficient assumption questions ONLY care about what if the assumption is true. Specifically, which answer guarantees the conclusion.
What makes some assumptions particularly tough is that sometimes, there’s an answer that would be right for the other type of assumption question (hence the popularity of answer choice (A) on “the rattlesnake question” (Google “LSAT rattlesnake question” if you’re unfamiliar with it). (A) would be right – if it were a sufficient assumption question. But it’s not.)
As Mike notes, sufficient assumption questions are much more likely to be based on connected-premises arguments, involving diagramming and conditional logic, so if a disproportionate amount of your trouble on assumption questions comes from sufficient assumption questions, that’s an important thing to work on. But the two biggest tips I can give you are –
1) Remember, an assumption is just an unstated premise.
2) Necessary assumption questions only want you to consider the implications of a false assumption (answer choice), and sufficient ones only want you to consider the implications of a true assumption. Therefore, while sufficient assumptions MAY also be necessary assumptions (and vice versa), it never matters whether or not they are.
Hope this helps.