Introduction

After 6 years, I took the GRE recently. As a quant-focused member on the team, I didn’t study for the quant (or the writing) at all, but did work on the verbal. Here’s what I can tell.

Disclaimer

I intentionally eschewed some of the suggestions Greg provided in his Must See videos. For instance, in Analyze Your Assumptions About Your Test Prep, we discourage students from creating their own study plan. I ignored it. I’ll explain why below.

And naturally, unless you know what you’re doing, ignoring an entire section of the test is a very bad idea.

Diagnostic

To start off, I took an official practice test (PowerPrep 2) to see where I stood at that time. I got a 159V and 168Q. I could tell immediately that vocab was hurting me significantly; when I first took the GRE in 2020, Greg himself reviewed my diagnostic report live and shredded my vocab knowledge (or the lack of it). When reviewing my test, it was clear that I got the SC/TC questions wrong primarily because I didn’t know the key words in the question, rather than an issue with logic or strategy. I also found weaknesses with my RC (reading comprehension) in the mid-160s range, but didn’t focus on that. The reason is that learning vocab is just memorisation and doesn’t require deep analysis - I was looking to get to the mid 160s, not a 170V.

This is why I didn’t follow a study plan and just did vocab, which is against Greg’s advice. The strategy part was “intuitive” to me and I didn’t find it necessary to watch the long videos separately.

This is an important strategy we regularly tell students. If you’re aiming for say a 165Q, it’s OK to not know how to do a few problems. What’s important is to nail down everything else. If I really wanted a 170V, I would have spent time on RC as well.

Preparation

Dogfooded the Vocab Mountain (and Vocab Check), some practice on the GregMAT site and from the Official Guide for verbal. I did 33 groups. Didn’t watch any videos. The main tip I’d recommend is to do something (a group) every day - discipline is pretty important here. Vocab Check is a useful tool as well - I got around 25 - 30 consistently.

One thing I'd like to note is the frustration I got when running into words I didn't know (and hence had to make them red). I could classify a word into three groups:

I knew the word from the start
I just needed to refresh my memory of the word (i.e, it's something I had seen before but couldn't recall the meaning of the word out-of-context)
I didn't know the meaning at all, or in a few cases had never heard of the word before.

Words in group (2) weren't that bad - within a couple of days at most I'd usually get them right. Words in group (3) were frustrating - so at times I only reviewed the words in red. Eventually some of the group 3 words become easier to remember; others were still stubborn and took longer to become green.

The extent of quant preparation was doing some random problems the day before the test, and for writing, watching this video.

Dogfooding refers to the practice of the members in a company using their products in real-life during development. In this case, it refers to my using Vocab Mountain (and Vocab Check) to actually improve my vocab for the GRE.

The actual test

My test was VQVQ (i.e, verbal came first).

Writing

Didn’t find the prompt particularly nice. Took a bit of time to think about what to say, and while I did follow Greg’s outline broadly, I felt like I struggled to elaborate on my examples within the time limit.

Verbal

In section 1, I didn’t find the vocab (and the test as a whole) particularly challenging, barring maybe one problem or so.

Section 2 was clearly harder, with the RCs requiring more time to think through and the logic for the TC questions being less straightforward as well. What I did notice was that I was able to cut through SC pretty quickly - Greg’s pairing strategy and quickly being able to recall vocab meant that even the harder SC problems in the second section didn’t feel difficult.

Quant

I just used autopilot and found the test straightforward; I thought some of the problems in the Official Guide were harder than what I got in the test (contrary to what I've heard people report on Reddit). I finished both sections within half time and then essentially twiddled my thumbs (almost). I’ll have more to tell in the next section.

Autopilot refers to just doing something (i.e, the test) without thinking. To give an analogy, if you're a new driver, you're likely to be cautious and careful whenever you drive, especially for the first few times. Conversely, if you've been driving for a while, would you need to "think" about how to drive every time you get into the car? No, it just comes naturally for you. No different here. Someone like Greg or Vince (or one of our verbal tutors on the site) would be able to autopilot through verbal as well; I can't.

I recommend every student to “take the test twice” (which is like checking your work on every problem), which is also what I did here. This helped since I found a silly mistake in a QC problem in section 1.

Reflection

I got a 162V/170Q/4 AWA. It would have been nice to get a couple of points more on verbal, but a 162V is acceptable. Same for the writing, but I didn’t study for it so I don’t think I could complain. Now let’s look at the diagnostic report:

Verbal

Let’s go back to my 2020 attempt. Out of the 8 SE questions in that test, I got a 2/8. Out of the 7 questions in this test, I got every single one of them correct. Notice the improvement which was just due to vocabulary. Conversely, I didn’t really do as well in TC as I hoped (7/12 vs 4/7) which was a disappointment as I thought that the vocab improvements would have directly translated to TC.

RC looks better in my second attempt (9/13 vs 12/20) but I am not sure I really improved there, especially given the shorter number of questions.

One thing I did observe is that hard ETS TC/SE questions work a bit differently from the TC/SC hard problems we have on our site. On our site, hard/extreme problems often involve difficult vocab (either in the word itself, or the usage will be very uncommon). In contrast, ETS hard problems tend to involve tricky logic instead of overly difficult vocab. On the other hand, I found our medium-difficulty problems to be pretty representive of ETS medium-difficulty equivalents.

Quant

First, I don’t really know why I got the difficulty 1 chart question wrong in the first section. I feel cursed; I have gotten one quant wrong on every official ETS test I’ve taken (practice or official). It is also important to not look at the “Time Spent” column seriously - the way I take the test is not the way most students would take.

Anyway, as I was taking the test, I was surprised at how direct it was. Out of the 27 questions, 26 of them could be mapped directly to PrepSwift videos - that’s how important foundation is. Some examples off the top of my head:

That Level 5 Numeric Entry question in Arithmetic (Section 2)? Similar to a PrepSwift video on percentages.
And the multiple select chart question in Section 1? One of the options was nearly identical to a percentage PrepSwift quiz question as well.
Reference #3 in Algebra (Section 2) combined the foundations of two PrepSwift videos on coordinate geometry.
The Numeric Entry problem in Algebra Section 2 used the exact same concept covered in a shortcut video on PrepSwift.
An Algebra QC problem in Section 1 (with a Level 3 difficulty) was again very similar to an Algebra video on PrepSwift. Looking back at it, I thought it was a level 1 problem - it was intuitive, simply tested the basics of that concept and nothing else.
The Level 5 Data Analysis real-life multiple-choice problem was directly based on a PrepSwift video on combinatorics.

There was one question that did surprise me slightly. It was the Level 5 QC problem in Geometry (Section 2). A spatial reasoning problem, it looked more like a brainteaser than a concept question per se. While we do cover the core ideas in PrepSwift Geometry, that specific question required a somewhat creative deduction that we hadn’t explicitly covered in the course (however: we do have these kinds of "surprise" problems on the timed quant sections and full practice tests). If I hadn’t seen a very similar problem before, I would have had to think a bit to work it out. Otherwise, I thought that the test was a bit easier than when I did it in 2020.

This goes against some students’ minds that the test has gotten harder over the years. I don’t agree. And even if the test did get harder, it’s not a bad thing (there are a whole lot of misconceptions which partially fuel these conspiracy theories). We’ll have more to say in a future blog post.

Strategies are also important, especially in verifying the correctness of a problem and whittling off incorrect options, but they are of little use without the necessary foundations. On that note, recall the multiple-select chart problem in Section 1. There were three options. It turned out that two of the options were identical, so something like this:

Option A
rephrased, the same thing as option A
Option C

I haven't seen this before in an officlal test. Notice that from 7 possible ways to select an answer, recognising that two options were identical brought the number of ways down to three.

Postscript

I didn't have my passport for a while and asked ETS whether I could take the test without it (as someone living in a country different to their nationality, a passport is ordinarily required). Turns out there is no point; exceptions are not generally granted to anyone other than refugees.