How to speed up without losing accuracy

[b91302310]

Under untimed condition (about 1.75 mins/question on average), my accuracy rate is higher (-3 ~ -5 for 25 questions in 45 mins). However, when I tried to complete more questions under timed condition (35mins), I had more questions incorrectly answered and those questions are not especially difficult. Even worse, some are quite easy ones which I can correctly answer under untimed condition. My accuracy rate is negatively correlated to the completion rate. So, how to speed up and at the same time, keep answering questions correctly?

Could anyone help??
Thanks.

[ManhattanPrepLSAT1]

You know, the truth is that to get faster you need to see these arguments very clearly and come to quick decisions as to where an argument takes a wrong turn. There are two other things you can do that I think are useful, but not sufficient.

1. pass over challenging questions. On questions where you've read several answer choices more than three times and still do not know the answer, leave them behind. You just cannot afford to get bogged down on one question. I sort of feel like I'm bouncing through the section on a pogo stick - and this dictates the rhythm and pace.

2. use your knowledge of where you're at in the section to frame your expectations for the answer choices. 1st third of the section = go with your first instincts and move quickly; 2nd third = be careful of shifts in language (ie: stimulus talks about kitchen appliances and answer choice talks about kitchen utensils) but a straightforward understanding of the argument is appropriate; final 3rd = use conceptual organization to understand the argument structure.

I use both of these tactics to keep my pace up in LR. Good luck!

[interestedintacos]

I've found your last point to be exceptionally useful. If the LSAT were like the GMAT then many test takers wouldn't even see as many easy questions as they end up seeing on the LSAT (often apparently in the first part of the section). The GMAT of course adjusts difficulty for the individual test taker based on whether he/she gets a question correct or incorrect.

I've found it's important to not think too hard about the easy questions (doing so can sometimes use up more time or even lead to getting them wrong), and conserve energy for the more difficult ones. However, is it always the case that the last part of the section will be the hardest? Has anyone done a statistical analysis on this? Why not plug in the numbers for each PT, including a ranking of difficulty (which is public info) for each question, analyzing where the level 1-3 questions are in a section and the level 4-5, for instance, over all 62 PTs. I've seen a lot of statistical analysis like that, but no data for this issue in particular.

[ManhattanPrepLSAT1]

It's not guaranteed that question 17 will be more difficult than question 7, but predictably it's nearly always the case.

If you'd like to get a sense for the fluctuation in difficulty, take a look at the Official LSAT SuperPrep which contains the Feb96, Feb99, and Feb00 LSATs. There are rankings for each question in terms of difficulty and if you track the changing difficulty, you can see that it's not as neat and tidy as one would hope, but that a generalization of the difficulty based on question location can be found.