Before the December 2016 LSAT I wrote an article attempting to predict that test’s content based on the material presented on the most recent LSATs, and in this post I’m going to examine those results and also make some informed guesses as to what the February 2017 exam may have in store. I’ll begin with the same intro that prefaced my previous post:
“The LSAT is not a static exam. Rather, it varies slightly from year to year and even from test to test, as LSAC explores the full scope of what’s allowable under the rigors of “standardized” testing. And while those metrics are indeed restrictive—introduce significant deviation, and the measurements themselves become unreliable, as I’ve written about previously—even the tiniest of tweaks can have a profound impact, turning a typical test day into a singular success…or something equally unique in the opposite direction.”
I then went on to discuss the idea of what constitutes true “change” and why it matters, but to spare you any redundancy I’ll cut right to the chase here and see how I did.
As always, LR variance is best measured by question (and concept, like Causality or Formal Logic) type. We’ve charted the question types by appearance count before, and that’s what I used as a baseline in my earlier post and what you should use as a rough guideline going forward.
That said, some notable trends were once again apparent. Let’s take a look at what I got right and wrong, and anything else that stands out.
Hits: Resolve the Paradox and Second Family questions in general. Also, Cannot Be True and Evaluate.
I noted in my earlier post that Resolve was being tested more frequently these days, and December did not disappoint. Despite a historical tendency of around 5-7%, the December test had five of these questions (as did September 2016), or about 10% (twice the average). I strongly encourage anyone preparing for February to master this type—and as mentioned before it can be quickly mastered with the right approach.
More generally, in December Second Family (Help) questions—which include Resolve as well as Strengthen, Assumption, and Justify—appeared 22 times, or nearly half of LR. I highlighted a recent emphasis on these types, and for those paying attention it paid off:
“Assumptions too have been creeping steadily upward in frequency, with a whopping eight on the September test! In fact, Second Family questions (Strengthen, Justify, Assumption, and Resolve) were somewhat surprisingly more common in September than First Family (Must, Flaw, etc). Granted, it was 24 to 23, but still that’s a crucial uptick and together those two Families dwarf Weaken and Cannot at just three questions total.”
There were only three Assumptions in December, but Strengthen picked up the slack, appearing nine times. I see no reason to think a continued emphasis won’t carry over to February.
Next, both Cannot Be True and Evaluate the Argument—two types that, until recent years, had been on the fringes of use—have continued to appear with some regularity. I mentioned pre-December that you should, “Expect to see at least one of each on the test,” and sure enough there was one Cannot (first LR) and one Evaluate (second LR). I expect the same for February.
Misses: Method of Reasoning-Argument Part.
These were extremely common on both the December 2015 LSAT and the June 2016 test (four to five on each, which is really heavy), but September 2016 only had two and December, surprisingly, didn’t have a single one. In fact, Method itself only appeared once!
So this is clearly the most unpredictable question type; what that means for February is hard to guess, but given the past year’s count I’d spend some time reviewing these just to be on the safe side.
Other items of note:
Flaw in the Reasoning and Must Be True, the two question types that have historically appeared most frequently, were once again marginalized on the December LSAT: only seven Flaw, and five Must questions were tested. In fact, the first section only had a single Must Be True question, and it didn’t appear until number 22. That’s highly unusual. And while it may not last long, it suggests that the Second Family types (mentioned above) likely deserve a higher priority at the moment.
Lastly, Principle questions—not a type per se, but rather a reasoning idea featured in a number of types—were a standout, with six instances between the two LR sections. Most of these were Strengthen-PR, but the frequent usage itself suggests it would be wise to brush up on Principle before February.
As I stated before, “While it’s always a fairly safe bet that you’ll encounter the faithful few in LG—Sequencing, Basic and Advanced Linear, Grouping, and (particularly of late) Grouping/Linear Combination—a savvy test taker is aware of four notable exceptions (and a general reemergence of oddities on the whole).”
I won’t go into all four of those outlier instances here (read the linked post already), but will point out that December 2016 most definitely contained a fifth!
Hits: Pattern Games.
It happened again. The fourth game in December was a (another) Pattern game. Here’s what I said before the exam:
“Pattern games appeared in each of the last two years—June 2014 and December 2015—but (unless reports from this past February are mistaken) not so far in 2016. I’d review the games from those two exams to be sure you’re prepared, just in case.”
Those who heeded my warning were at least, in some small measure, equipped to tackle game 4. The rest? Let’s just say it wasn’t pretty. So, I’ll repeat: spend your time on the “guarantees,” (Basic Linear, Advanced Linear, Grouping), but don’t get complacent once you have the basics covered. Chances are better than ever that the test makers could throw you a curve ball! Be ready for it.
Other items of note: Numerical Distributions.
September and December both saw a heavy emphasis on numbers in games, with “mismatched” (uneven) variable sets when comparing the selection group to the base. I suspect February will continue this focus. Be sure to practice with games featuring imbalance before test day!
I’ll once again begin by repeating my preface from the prior article:
“This is perhaps the most challenging section for “trends,” since both difficulty and outliers are hard to objectively quantify. However, consensus among many seems to be that RC is getting harder. Take the Eileen Gray (Lacquer) passage from September 2016 as just one example: this was almost unanimously considered the toughest passage on that test and, while I’ll admit this is anecdotal, it caused about as much online outrage as I’ve seen in recent memory.
So I recommend reviewing not only that passage, but any that have given you particular trouble on recent (recent-year) practice tests. Odds are you’ll see at least one of similar difficulty on the next LSAT.”
Rather than do a “Hits” and “Misses” breakdown, I’ll sum up my prediction success more succinctly: the December 2016 Reading Comp was, by nearly all accounts, tremendously benign (certainly much more so than past tests led me to anticipate). So we can call this one mostly a miss, but one I’m sure no one is too upset by—if you over-prepared for RC with notoriously hard passages you no doubt found December generous, and if you were under-prepared you got off easy. Be careful though! The overall trend is still one of escalating difficulty, and February, more likely than not, will hold true to that.
So that’s my December breakdown, both in terms of what I suspected and what it all (likely) means. I’ll wrap up with the same sentiment I used to conclude the previous discussion: “the sectional analyses above are all about recent points of emphasis and notable trends, and should be treated as such. There’s no guarantee that LSAC stays the course and repeats these patterns, but by using their behavior to inform yours you give yourself the best possible shot at a thoroughly predictable test day.” Good luck in February!
Image: “elevator” courtesy of Gideon Tsang.