Getting ready for the LSAT? Use my day-by-day study plans and free LSAT videos to stay on track.

Save time by instantly downloading LSAT PrepTest PDFs and explanations.

* PrepTest 73 (the September 2014 LSAT) is now available for instant PDF download! *

April 2, 2010

The LSAT is Hard, Easy, Bullshit, and Unfair

LSAT Blog Hard Easy Bullshit 2Of course, I don't believe it's bullshit or unfair.

Apparently, you guys do. Google told me so.

According to Google's search suggestions (based on billions of searches), you think it's hard, easy, bullshit, and unfair, among other things:



LSAT is hard? Agreed.

LSAT is easy? Agreed.

LSAT is bullshit and unfair? Nope. LSAC says it has a correlation of .35 with 1L grades, and that's an underestimation because some test-takers never make it to law school (see p16 of PDF).

LSAT is out of...control? It's out of 180.

LSAT is tomorrow? Not usually. See LSAT dates.

How the LSAT is graded? I did a big series on the LSAT "curve" recently.

Which LSAT is the easiest / hardest? No particular LSAT exam is easier or harder than any other at the end of the day. Again, see my series on the LSAT "curve."




9 comments:

  1. Can I even tell you how excited I am that you used "bullshit" in a post? I fracking love it. Please use "frack" ASAP.
    <3 <3 <3

    ReplyDelete
  2. don't use frack

    ReplyDelete
  3. Do you really consider to be .35 such a great correlation? That is closer to 0 (no correlation) than 1 (perfect correlation.

    ...?

    ReplyDelete
  4. To the first 2 commenters:

    Haven't decided whether I'll use "frack" yet, haha.

    ***

    To the third commenter:

    I don't think it's a great correlation, but it's not terrible. LSAT alone actually has stronger correlation with 1L grades than undergrad GPA alone.

    The same PDF states that undergrad GPA only has a .28 correlation with 1L grades. A composite of LSAT and undergrad GPA has a .47 correlation with 1L grades. (This is all on page 16 of the PDF.)

    I don't know of a better predictor of 1L grades than this composite. Please let me know if you find one, and I'll work on selling it to LSAC and the law schools.

    ReplyDelete
  5. Hi Steve

    Thanks for creating this blog. It was super helpful for me as I prepped for my February test. I've read your posts on retakes, yet I still have a question regarding retaking the test (which I plan on doing in June): how should I approach it? Since February was an undisclosed test, I really don't know which question types/section I did well on or flunked on. I don't think I was particularly stressed or under-prepared going in either. Aside from going through all the PTs again, I don't know how I can study differently. It seems like I'll have to take a blanket approach again ... and well, I have the nagging feeling that I should do something /different/ this time around.

    Any advice will be much appreciated.

    ReplyDelete
  6. Hi Tonka Time,

    Since Feb tests are undisclosed, take a practice test to see where you stand, then use that to jump on the retakers' schedule.

    Yes, you should do something different.

    -Study more (if you didn't study enough)
    -Study less (if you burned out)
    -Study differently

    Check out my regular 2-month LSAT study schedule if you didn't use it the first time around.

    ReplyDelete
  7. >>>and that's an underestimation because some test-takers never make it to law school (see p16 of PDF).

    What? How can we possibly know how the LSAT correlates to law school grades no one ever received? For all we know, these non-law school students would have not affected the correlation or would have weakened it.

    ReplyDelete
  8. Excellent point, Terry! Without more information, we have no way of knowing.

    However, if you look at p16 of the PDF I linked above, you'll find this gem:

    Because the data derive from a restricted sample (i.e., those who are actually selected for law school), they underestimate the validity of the combination. Dalessandro et al. (2005) demonstrate that correlation coefficients are higher when the data are based on an unselected sample.

    ReplyDelete
  9. Steve, I had looked at the page prior to posting, but I couldn't see how Dalessandro or anyone else could make that deduction. He may have an *argument* that that's the case, but I can't imagine how he could prove it. Anyway, first day with your blog, quite nice.

    ReplyDelete