LSAT Curve

shegotit

I see a lot of discussions of the LSAT having a curve but what exactly determines how much the curve is worth. What exactly is the curve and what is it based on?

Leah M B

The "curve" of a test isn't really a traditional curve. What people mean is, the most questions you can miss and still get a 170. That can vary from -9 to -14. So if a test is a -14 curve, that means it was significantly difficult and you have a little more leeway to miss questions and still get a good score. If it's -9, that means the test was on the easier side and there's very little slack. Most tests are more in the -11 to -13 range. But this is the way that LSAC adjusts for some differences in the tests, and what makes them all pretty equal with each other as far as your final score goes.

youbbyun

@"Leah M B" do they set the curve based on the experimental, so the curve is set before the exam is even given -- or its the curve set AFTER the exam takes place, taking into account how well ppl did on the exam?

samantha.ashley92

My understanding is that the curve is generally based on a true bell curve. Assuming you didn't go to college in Canada, when we think about "curving" grades in college, we are usually thinking about grade inflation. That's not how the LSAT works. Based on what I've seen, the bell curve is set so that a 151-152 winds up being the 50th percentile, 160 is around the 80th percentile, and a 170-171 is the 98th percentile.

Depending on how many questions you get right, this could be really helpful. Alternatively, it could kind of screw you over. In a completely unrealistic situation, let's say that 50% of people get 60-67 questions right. If you score in that category, every single question is going to be hugely influential on your score. Otherwise, it would be crazy difficult to create a bell curve. (Aka you'll probably get screwed over.) That also means that it would be even more rare than usual that people got 95+ questions right. Or maybe twice the average score 95+ questions right. Stuff like this is why some tests don't include a certain score at all. For example, PT 70 does not allow you to score a 178. You can see a * for 178 on the PT's conversion chart. Does that make sense?

Leah M B

@"samantha.ashley92" I'm not sure about the details of all the scores like that, but generally when people talk about the curve on the forum here, they are referencing that number I mentioned, how many you can miss and still get a 170.

@username_hello Honestly, I don't know. Haven't dug that much into it. I just know that it's different for every test. If you haven't checked out Dave Killoran's blog, it's really informative and I wouldn't be surprised if there's a post that answers this somewhere. https://blog.powerscore.com/lsat

eRetaker

It's determined by the experimental sections before the test is even administered. Though there are rare instances where LSAC had to adjust a curve after the exam is taken.

akistotle

All I know is that the LSAT isn't actually scored to a curve. What I mean by that is: The LSAT isn't curved on how test takers perform on the day of one LSAT administration. Taking the LSAT among a lower-scoring pool does not give you an advantage.

It goes through a statistical process called test equating.

Raw scores are converted to an LSAT scale that ranges from 120 to 180, with 120 being the lowest possible score and 180 the highest possible score. This is done through a statistical procedure known as equating, a method that adjusts for minor differences in difficulty between test forms.

https://www.lsac.org/jd/lsat/your-score

Pre-equating is a statistical method used to adjust for minor fluctuations in the difficulty of different test forms so that a test taker is neither advantaged nor disadvantaged by the particular form that is given.

https://www.lsac.org/jd/lsat/policies/challenging-lsat-questions

Basically, you don't have to worry about it too much

MichaelScottRegionalManager

@akistotle I think what the website says is that the level of difficulty for each test is predetermined so that no one test will be harder of easier than another. For example, if they have a trash logic games section (like the one we had last Saturday) then they’ll make up for it by making the logical reasoning less difficult.

I still think there’s a “curve” based on our performance after we take the test. Like everyone is saying, it’s not really a curve since it’s there’s not a set number of people who will score 170 or anything like that. But there is some leniency offered for questions that everyone missed. For example, if everyone missed a particular question, then they’ll “exclude” that question to give a -11 curve. Of course there will be other factors taken into consideration, but I’m sure our actual performance on the test plays a role in determining the curve.

I called LSAC just in case (because I’m super ptsd’ed about Saturday’s logic games section) and they said that the curve will take into account all the canceled scores as well. This means that we don’t have to worry about whether it will only take into account a flawed sample of those who didn’t cancel.

I have no idea if I’m making sense, but I hope this helps others like me who spent the last two days sulking...🥺