PDA

View Full Version : So I took the GRE yesterday, and the analytical section has gone missing.


Spectre of Pithecanthropus
03-31-2006, 11:59 AM
So yesterday I took the GRE General Test. I was held up by work problems until much later than I should have left the house, and then I started threading through the West L.A. midday traffic to reach the testing center, located in the Culver City Panhandle on West Slauson Blvd. There didn't seem to be addresses on any of the buildings, so I ended up asking a mail carrier who was taking a break in his truck, eating lunch. Surprisingly enough, it turned out to be in a business park where I worked for a year, about nine years ago.

I didn't feel ready for the test. Other obligations, mostly work related, seemed fated to obtrude themselves into mornings and afternoons, here and there, that I had set aside for studying. Between that and being flustered by the time I arrived at the center and went through the preliminaries.

I sat down in the cubicle that had been allotted to me. Deliberately, I went through the tutorials on how to answer the test questions with the mouse. It was totally unnecessary for me, of course, but I was actually hesitant to jump over the edge and begin the first timed section.

I worked through the test--first, a new section on analytical writing. Arrgh! I haven't had someone assign me a topic to write about in 20 years, and even then I was a graduate student being asked to write about a topic in my field. Next came the quantitative section. I have never finished the quantitative section, but I did get nearly all the way through. And the verbal section, always my strong point, threw me a few zingers as well. (Note to self--learn and remember what "ineluctable" means). Some of the analogy questions didn't seem to make a whole lot of sense. But I made it through, and was all set for the next section, the analytical, when there was....

the end of the test. My quantitative and verbal scores, which are now instantly displayed and reported came back and I couldn't believe it. As badly as I though I'd done, and with as little preparation as I'd had, my verbal was only 20 points less than the last time, and my quantitative was actually 10 points higher than before. The analytical writing section will be manually scored by readers and I'll get the results in a couple of weeks.

But what happened to the old analytical section--the one that used to have logic puzzles and so forth? Does the writing section replace it? I'm concerned because some preliminary Googling on the GRE suggests that the analytical section is still supposed to be there.

Shagnasty
03-31-2006, 12:06 PM
They just chucked it. That section was never very successful and a lot of grad schools didn'y really know what to do with the results of that section. I believe the results skewed into odd distributions as well (lots at the top) and that isn't good for that sort of test.

They decided a writing section would be more useful for grad schools that use the GRE.

That is it.

Anne Neville
03-31-2006, 12:12 PM
But what happened to the old analytical section--the one that used to have logic puzzles and so forth? Does the writing section replace it? I'm concerned because some preliminary Googling on the GRE suggests that the analytical section is still supposed to be there.

They got rid of it in October 2002 and replaced it with the analytical writing section. (http://en.wikipedia.org/wiki/Graduate_Record_Examination#GRE_prior_to_October_2002)

You gonna tell us how you did?

Spectre of Pithecanthropus
03-31-2006, 12:30 PM
They got rid of it in October 2002 and replaced it with the analytical writing section. (http://en.wikipedia.org/wiki/Graduate_Record_Examination#GRE_prior_to_October_2002)

I thought it must be something like that. Google is much, much better than the early search engines, but still doesn't weed out results with outdated information.



You gonna tell us how you did?

What the hell: 700 verbal 690 quantitative. You can deduce my prior scores from my statements in the OP. From the information I've gleaned, the verbal section is more difficult in terms of how the percentiles work out. At any given score between 200 and 800, many more people attain that result on the quantitative than they do on the verbal. This makes sense, since the quantitative section, IMO, is mostly high school math, while the verbal section definitely does resolve around college level vocabulary.


Oddly enough, that last time, my quantitatve and analytical scores equalled each other.

Anne Neville
03-31-2006, 01:01 PM
What the hell: 700 verbal 690 quantitative.

Congratulations! That's quite good.

From the information I've gleaned, the verbal section is more difficult in terms of how the percentiles work out. At any given score between 200 and 800, many more people attain that result on the quantitative than they do on the verbal. This makes sense, since the quantitative section, IMO, is mostly high school math, while the verbal section definitely does resolve around college level vocabulary.

I had always heard about people doing much better on the quantitative than the verbal section, but I thought that was a selection effect because of the sort of people that I'm likely to discuss GRE scores with- I was a science major, and my coworkers now are all computer people. Those types of people always do tend to do better on quantitative tests than on verbal ones.

I'm kind of the exception to that, because I'm terrible at arithmetic. I was a math major (among other things) and did pretty well with higher math, but I have some mental block so that I can't do arithmetic anything like quickly or accurately in my head. My GRE quantitative score was higher than my verbal score (770 versus 730- you told me how you did, I might as well tell you how I did), though it was reversed for my SAT scores (~660 math versus ~720 verbal- my memory's a little fuzzy on those, but I think that's what they were).

Before I took my GRE, I had nightmares about it several times. Usually they involved me driving to take the test (which was odd, because I didn't even have a driver's license then) and not being able to find it. I was driving around and around, and it was getting later and later. I could see the building where I was supposed to take the test off in the distance, but no matter what I did, I couldn't get any closer to it. Those nightmares freaked me out so much (plus there's the fact that in my waking life, I have a lousy sense of direction) that I ended up going to the place where I was going to take the test beforehand, just to make sure I knew how to get there. I timed the trip, too, so that I knew how long it would take.

Spectre of Pithecanthropus
03-31-2006, 01:03 PM
while the verbal section definitely does resolve around college level vocabulary.

.

Cripes, revolve around college level vocabulary.

gigi
03-31-2006, 01:19 PM
I took the GREs years ago, and when a friend took them more recently I was a little shocked that they got rid of analytical and replaced it with writing! How can they score writing objectively? What about crappy typists? Of course I'm also weirded out that there is no written option; I like being able to work on paper and mark up the answers on paper right next to the test, rather than looking up and down.

Plus I did well on analytical, and could tell which part was experimental and wouldn't count. (V-720, Q-800, A-800, but I always was disturbed by the fact that I got questions wrong and still got a "perfect" score :confused: )

Spectre of Pithecanthropus
03-31-2006, 01:39 PM
Plus I did well on analytical, and could tell which part was experimental and wouldn't count. (V-720, Q-800, A-800, but I always was disturbed by the fact that I got questions wrong and still got a "perfect" score :confused: )

Damn you're good!

The computer adaptive tests modify the questions answered based on how you are doing. If you're doing well you get harder questions, which fact, I think, gets factored into the score computation. If you took a CAT test, that might be the reason you managed perfect scores notwithstanding a few wrong or skipped answers.

I know I though I'd done just terribly on the quantitative, and I'm sure there were some wrong answers there, as well as the ones that I didn't get to, but the result was a pleasant surprise. I didn't reach average engineer or physicist level on the quantitative section, but I came close.

Spectre of Pithecanthropus
03-31-2006, 02:04 PM
Congratulations! That's quite good.

You're too kind.

I had always heard about people doing much better on the quantitative than the verbal section, but I thought that was a selection effect because of the sort of people that I'm likely to discuss GRE scores with- I was a science major, and my coworkers now are all computer people. Those types of people always do tend to do better on quantitative tests than on verbal ones.
.

It's definitely a verified result. I remember that there was similar divergence nine years ago, but I don't think it was as stark as this then. For instance, this link (http://www.cbtkorea.or.kr/download/01210.pdf) from 2003/04 indicates that a verbal score of 700 puts you in the 96th percentile, while an equal quantitative result puts you only in the 70th percentile.

Mathochist
03-31-2006, 09:40 PM
They just chucked it. That section was never very successful and a lot of grad schools didn'y really know what to do with the results of that section. I believe the results skewed into odd distributions as well (lots at the top) and that isn't good for that sort of test.

They decided a writing section would be more useful for grad schools that use the GRE.

That is it.

Of course, for mathematics graduate study, the situation is exactly reversed. But screw the mathematicians. They've never done anything as good for society as a lit-crit Ph.D., have they?

Shagnasty
03-31-2006, 10:24 PM
Of course, for mathematics graduate study, the situation is exactly reversed. But screw the mathematicians. They've never done anything as good for society as a lit-crit Ph.D., have they?

No seriously, I know more about psychometrics than the average Joe and the old analytical section had serious problems.

Ideally what you want in such a test is a nice normal curve. Most people will fall in the middle and you can test way out into many standard deviations. If someone gets a 700 on a subtest with a mean of 500, there is still room for them to score higher but they still tested very far above the norm. Normal curve charts can tell you how that plays out.

The verbal section does that. People follow the normal curve and it is extremely difficult to approach the 800 mark. That is done by only one person in several thousand. That is good because you have test that distributes people well on a measurable scale and leaves room for unusually high scores.

The quantitative test is where you should be focusing your energy. This test has problems in theory but it isn't the fault of the test. You see a bimodal distribution where the distribution shows a normal curve except it has two humps like two waves right behind each other. It is showing that there is one group of people that sort of get quantitative thinking and that builds on a curve and the goes down and then it builds again and there is another crest that shows where most of these people fall. A good number of these people ride the crest of the wave all the way to the end of the chart.

This is not good because tests need to differentiate one another. You want to provide sufficient difficulty that you can show that this person is so good in math that only 1 in 10,000 can score that high. However, you can't because you have to measure everyone and there is a whole distribution that is so far behind the pack in front that it becomes hard to measure. It is possible to design such a test to figure out everyone's level but it takes time and money possibly on the test taker's as well as test designers. They tend to use the optional subtest in math for that.

Now, the old analytical section was all FUBAR. I had a work study job as an undergraduate processing graduate admissions applications. I had to input GRE scores into a database and the old analytical score was not included in their data. My college job doesn't sound very important but I was there with the graduate committee the whole time for issues and problems they had. They said they had no research to back up what an old analytical score meant and they saw far too many people that scored fairly low on the other subtests and very high on the analytical section.

I took the GRE and was exhausted by the time I got to the analytical section. I decided to not do it at all. That's right, I just thought and doodled the entire section and randomly filled in bubbles. When my scores came back, the others were good and the analytical was predicatively bad. I just noted that I decided not to take that section because it has no relevance on my applications. I got into my first choice Ivy League school (Dartmouth) and that was that. Professors there told me they couldn't care less about an analytical GRE score and they had no idea what to do with it.

The problem with the old test is that it didn't correlate well with the other scores. Fans of multiple intelligences may see that as a bonus but nobody could ever figure out what it was measuring. People say the same about IQ tests but those are statistically rock-solid and certainly measuring something even if we don't know what it is exactly. A bigger problem is that a huge number of people got perfect scores on the analytical section. Again, that is terrible from a testing standpoint. The test needs to differentiate people and have people know roughly what that means.

I took the new GRE two years ago. I love the new computer format and find it less stressful. Computer Adaptive Tests are sound scientifically and make for better test taking as well. I would rather a system try to fine tune my score around a point rather than take a massive number or items that are either too easy or way too hard. The instant results are by far the best part. I did well too although not as well as some of you. I am always puzzled when people say standardized test don't measure anything. They may not measure things relevant to some things but they certainly measure something. I had my first IQ test at 6 years old. I have taken literally dozens of casual and professionally administered test since then. The score is always roughly the same and within a very small percentage point difference. I don't even know why they bother. There are conversion charts for all major tests and they are oddly accurate.

Shagnasty
03-31-2006, 10:34 PM
Of course, for mathematics graduate study, the situation is exactly reversed. But screw the mathematicians. They've never done anything as good for society as a lit-crit Ph.D., have they?

In case my detailed explanation was too long, it seems you have no idea what the analytical section was. It was a games type section and the mathematical section is called the quantitative section, not the analytical. There is also a math subtest which measures advanced math ability in much more detail. I went to graduate school in a hard science after taking it (behavioral neuroscience)

I do think the verbal section is important in all fields because it can weed out people with substandard reading comprehension and that is an issue that seems to pop up frequently.

bitwise
03-31-2006, 10:38 PM
Of course, for mathematics graduate study, the situation is exactly reversed. But screw the mathematicians. They've never done anything as good for society as a lit-crit Ph.D., have they?

Actually, I think training in mathematics made me do worse on that section. Part of the reason I did poorly was that I didn't study. However, I think mathematics trained me to be prejudiced against the type of questions asked on the test, which basically require you to evaluate the truth of certain statements from arbitrary conditions with no context. I kept wanting to rearrange the conditions into a cleaner form.

The worse questions were of the form "Which of the following statements could be true?" I think for some questions like this, I wasted a lot of time trying to prove one of the independent statements false.

Spectre of Pithecanthropus
04-01-2006, 12:15 PM
Of course, for mathematics graduate study, the situation is exactly reversed. But screw the mathematicians. They've never done anything as good for society as a lit-crit Ph.D., have they?
This is uncalled for. There's no reason this should devolve into a humanities vs. science slanging match.

Besides, isn't there a subject GRE for math? Maybe that's why the quantitative section skews so much "easier" than the verbal...if you're planning on grad study in physics, math, or engineering, you usually have to take an appropriate subject GRE.

Eureka
04-01-2006, 01:40 PM
To me, the quantitative seemed easier (and less frustrating to study for) than the verbal because if you understand the mechanics, all that changes are the numbers. If you can't do simple math quickly, you may have difficulties, or if you reverse numbers or swap the order of operations you may get answers that look right but are actually wrong. Guessing should never be neccessary.

With the verbal section, if you get a set of words that you have never seen before, you may just have to guess which one is correct. Sure, having a bigger vocabulary helps, but after a certain point the likelihood that any new word you try to study will show up on the test becomes so small it isn't worth the effort.

The analytical section was kind of fun, but the new writing section isn't difficult either.

Signed, A person who took the GRE twice, once on paper a while back, and once the computer version a couple years ago. I don't recall my paper scores, but when I took it on the computer, I got well over 700 on the Verbal, and exactly 800 on the Quantitative.

Shalmanese
04-01-2006, 05:55 PM
If you download the GRE Powerprep software, it contains all the data for distributions of GRE scores and allows you to drill down by test, profession and degree.

Mathochist
04-01-2006, 09:40 PM
In case my detailed explanation was too long, it seems you have no idea what the analytical section was. It was a games type section and the mathematical section is called the quantitative section, not the analytical. There is also a math subtest which measures advanced math ability in much more detail. I went to graduate school in a hard science after taking it (behavioral neuroscience)

I know exactly what it was, and how it dealt more directly with abstracty reasoning ability (given p, which of qn must/might/might not/can't be true) than the piddling math section ever could. That sort of thing is much more useful to a mathematics department than the high-school test the rest of it is.

CookingWithGas
04-01-2006, 10:05 PM
I am always puzzled when people say standardized test don't measure anything. They may not measure things relevant to some things but they certainly measure something. . . . .[My] score is always roughly the same and within a very small percentage point difference. I don't even know why they bother. Who are the people who say that? I agree they measure something but it's not clear whether it's at all useful to measure it, or even whether anybody really knows what they're measuring.

I think a more common statement is that they don't correlate to real-world performance. (I am not claiming it's true, only that I have heard it many times. I do not have statistics to back that up.) I did very well on standardized tests in high school, but did poorly in college (too much beer, women, and guitar playing, often all at the same time). You can't generalize from that but it would be interesting to look for correlation of test scores to some measure of success, like income within a chosen profession (normalized for variables like geography).

Spectre of Pithecanthropus
04-02-2006, 12:00 PM
Actually, I think training in mathematics made me do worse on that section.

In hindsight, I think the quantitative section should not be regarded simply as a math test, although it looks a lot like one, and math and science folks usually do well on it. It's more a test of your ability to look at the problem given and quickly change your perspective or figure out the needed principle, without going beyond what they're asking for. For instance, in my recent test there was on problem involving two simple equations, but with three variables. For a couple of seconds I fumed silently--I had already been annoyed and flustered when I got there--then I realized I wasn't being asked to solve the equations, and that there was enough information there to derive the answer. IME that's not really being "mathematically inclined", but just using good sense.