Will new GRE percentiles be skewed?

[crater21]

Several people I know who have taken the new GRE have reported unusually high score ranges (based on the old scoring range). They weren't even expecting to do that well, which makes me wonder if the new GRE percentiles are going to be skewed when they are released in November.

It makes sense that scores are going to be higher. In the new format, people can go back and change answers, they can skip questions within a section, there is a calculator, etc.

So, if scores are higher on average in the new GRE, does this mean that even a relatively high score could fall under a lower percentile? What if you end up doing well in terms of the score, but your percentile is low? Will it affect admissions? How will schools interpret the new scores?

I'm just thinking out loud here, and could be totally wrong. But, if you have any thoughts about this, I'd be interested to hear them.

[Insei]

It's always possible, but if the average score (much like quant scores on the old exam) is heavily skewed, the validity of the test will be called into question. If ETS screws up this revision badly enough, maybe the GRE will cease to exist. Honestly, would anyone miss it?

The score I got on the GRE was on par with the last few practice tests I completed prior to the exam, so I can't personally speak to an unexpectedly high score. The people you talked to are a very tiny percentage of the hundreds of thousands of people who will take the test before November, so hopefully you've just got a biased sample.

[cunninlynguist]

It does seem like a lot of new scores are high, but ultimately the percentiles will always carry more weight. If you did well, though, I imagine the percentiles will be solid. After all, the new test has only been in effect for 2 weeks - perhaps the estimated scores are somewhat inaccurate (the 100-point range doesn't grant much confidence).

[Insei]

Another possibility, though purely speculation, is that the people taking the new GRE during this incipient period are doing so because they preferred (i.e. achieved better scores with) this new test format based on available practice tests. Additionally, one could argue that the folks taking it right now were fully prepared for the test change and chose to tackle it directly rather than blindly stumbling into it.

This accurately describes my decision to take the new test, anyway. That and the 50% discount, of course

Edited August 17, 2011 by Insei

[Dylan_Ireland]

Perhaps we should set up a poll here in order to gain a sense for the percentiles?

It would be a biased sample but it could be useful.

[chak]

or maybe the individuals with low test scores are: embarrassed to post it and/or don't practice as much as the individuals that use tools like thegradcafe forums to strategize their application process to get into graduate school.

Edited August 17, 2011 by chak

[crater21]

Yep, those are all very valid points.

I'm sure the experience of a few people is not representative. I was just thinking out loud because it struck me as a little surprising that almost everyone I talked to who's written the new test got very high score ranges.

I know one person who got 980 combined on the old test. And, in the new test she got score ranges of 720 - 800 on both sections. That's almost a 500-point jump. Now, maybe she just studied really, really hard. I don't know. But, she herself was surprised, and didn't think she had done all that well, and had in fact left a few Quant questions blank.

Anyway, I guess we will find out in November. My concern was basically that the percentiles are going to be really skewed. Even a really high score might not guarantee you a high percentile, which could potentially hurt you in admissions. That's what I'm worried about. I'm kicking myself for not taking it before the new format was introduced!

[orst11]

Maybe the fact that this is thegradcafe makes a difference in scores we are seeing. I note this because I have heard some people elsewhere saying they thought aspects of it were harder than the old version. I also personally know people with ranges far below what I have seen posted here.

[sollee]

Agree with all the comments people have made about various reasons we might be seeing higher scores here - mine were right about where I expected them to be based on the one ETS practice test I took a few days before.

Edited August 19, 2011 by sollee

[crater21]

Hi - sorry, I hope I'm not worrying people unnecessarily, so my apologies in advance. But, the practice tests are also based on the new test, aren't they? So, if the scores are high, then they would be high on the practice tests as well.

I took the practice test today, and I got a range of 720 - 800 on both sections. This was pretty much a shock to me because I can tell you that I am NOT a math person AT ALL. I'm a humanities student, who hasn't done math since high school. I also left about three or four questions blank on each of the quant sections because I ran out of time. So, yeah, this does nothing to allay my concerns about scores being unusually high.

The only explanation I can think of is that when they translate these scores to the new scale (130 - 170), it is possible that a high score on the old range may translate into something lower on the new scale.

[bentforyoga]

The thing is when they translate the scores, they still use data from the previous 3 years of testing, which I saw on the GRE site today after some clicking around. It wouldn't be fair would it to just use results to scale based on testing results from August 1? So then maybe the old scores would just translate to a seemingly lower 'looking' score on the new scale? Or at least until everything gets a bit more familiarized with the scale and the averages.

[wildviolet]

The score range I received on the verbal section was almost exactly what was predicted by the practice tests. Out of 20 questions, I would miss only 3 or 4 questions. I did better than I thought I would on the quantitative section, but I also thought the real exam was easier than the practice tests.

[H24]

I'm a bit worried the initial idea of people scoring high might be right...i took the actual revised test last week and got a total range of 1260-1460. That was 620-720 on verbal and 640-740 on math.

Two years ago when i took it to get into my master's program, i got an 1120 (550 verbal, 570 math). I studied a ton this time, probably about 3-3.5 weeks at 40 hours a week, if i added it up into regular work week format (though it was actually spread over about a month and a half); this was a good bit more than i studied last time. But that kind of increase does make me nervous....

As i posted in another topic, there's also the whole other issue of what this range of mine is actually equal to given the fact that i took the new test (i.e., a 1300 on the old test is not the same, ostensibly, as a 1300 on the new test)...i hate this.

Edited September 1, 2011 by H24