I actually love the simplicity with which the previous poster answered. There's no way I'll be that concise.
Keep in mind that Assumptions really perform 1 of 2 functions, so you predict them in different ways.
FUNCTION 1
Provide a missing logical link Sometimes that missing link feels like a "term shift"; other times it feels like a "missing premise"
There's no functional difference there, just a different feel.
"Term shift"
Caroline likes to make people feel good about themselves. Thus, Caroline encourages arrogance in other people.
"Missing premise"
Caroline likes to make people feel good about themselves. Thus, Caroline would make a good politician.
The assumption to each of these is the same ... "making people feel good about themselves [encourages arrogance] / [contributes to being a good politician]"
The way to get better at seeing these is to match things up as you read. In those two examples, 'Caroline' is a matching idea, i.e., it's mentioned twice. Anything mentioned twice in the argument won't need to appear in the correct answer. We're looking for the things only mentioned once.
Check out this random Sufficient Assumption question I pulled from PT26, Sec. 2, #21
The companies that are the prime purchases of computer software will not buy a software package if the costs of training staff to use it are high, and we know that it is expensive to teach people a software package that demands the memorization of unfamiliar commands. As a result, to be successful, commercial computer software cannot require users to memorize unfamiliar commands.Which of the following, if assumed, proves the conclusion?
Here's how I would match that one up:
The companies that are the prime purchases of computer software will not buy
a software package if the costs of training staff to use it are high, and we know that
it is expensive to teach people a software package that
demands the memorization of unfamiliar commands. As a result, to be successful, commercial computer software cannot
require users to memorize unfamiliar commands.What are the loose ends that are only mentioned once?
The companies that are the prime purchases of computer software will not buy and
to be successful, commercial computer software cannotCorrect answer?
Commercial computer software will not be successful unless prime purchasers buy it.
So Assumption questions that are testing missing links are looking for an answer to join together two ideas that were each only mentioned ONCE in the argument.
SHORTCUT: if there is a word/idea that only appears in the conclusion, then it must be in the correct answer as well on an assumption question.
The conclusion of that above argument was "As a result, to be successful, commercial computer software cannot require users to memorize unfamiliar commands."
The new wording was "to be successful, commercial computer software cannot".
Only two answer choices of the five have wording about what it would take for commercial computer software to be / not be successful. So only those two choices are really worth reading.
FUNCTION 2
The other way that a correct Assumption answer choice works is by ruling out an objection.
These are not explicitly predictable the way the "missing link" answers are.
These are more "fuzzy" predictions based on the type of argument.
- Author trying to say that X explains Y
(assumes that X could possibly cause Y and, more importantly, that something else did not cause Y)
- Author makes an analogy
(assumes that the two parts of the analogy are fair to compare in the way the author intends)
- Author tries to offer a solution to a problem / means to an end
(assumes the solution/method is a viable way of achieving that result without somehow backfiring and thwarting the goal)
- Author interprets a statistic
(assumes that there isn't any other way to interpret that statistic)
- Author disagrees with someone's claim
(assumes that the author's reason for objecting is in fact evidence that the original claim was wrong)
- Author infers something from a Sample / Study / Survey / Experiment
(assumes that the methodology of the evidence was sound ... representative sample ... no bias that would skew the data)
And so forth ...
For FUNCTION 2 assumptions, you're not predicting "a gap", but rather preparing your brain for how you (and LSAT) would object to the given type of argument.
Hope this helps!!