Page 1 of 1

How can I break 600 on this test?

Posted: Sun Nov 07, 2010 6:39 pm
by Blue Slime
The physics GRE is the bane of my existence.

I'll be taking it for the 3rd time next Saturday. My score from last year was 560, and all my applications were rejected. This time I had much more time to study since I'm no longer working, but despite filling up a notebook and a half with about a month and a half of daily study I haven't been able to break 590 on any of the 4 practice exams. Between this year and last, I've taken each of the practice exams twice, and reviewed the results carefully.

Now I have one more week left and it is looking very much like this years test (and therefore applications) will be a repeat of last year. Does anyone have any golden advice for this test in such a last-minute situation?

My goal for this year was to get a 650, which would at least put me over the 50th percentile. By now I'd just be happy to break 600 and hope just getting out of the 500's would give my application a better chance.

From my own analysis, my biggest problem is being pressed for time, and not knowing when to stop when I'm not making progress on a problem. In the test of Nov 2009, I finished without even seeing the last 25 questions. I've mitigated this a bit by having a strategy where I pass through the test once only doing problems which require no calculation, marking the ones I think I could solve with some calculation for later. Then on the second pass its all calculations. This way I at least ensure I see all of the questions, and don't miss any easy ones.

My statistics for this year:

Test | Answered | Correct | Raw Score | Scaled Score
GR8677 | 77 | 40 | 30.75 | 590
GR9277 | 68 | 35 | 26.75 | 580
GR9677 | 45 | 18 | 11.25 | 530
GR0177 | 52 | 33 | 28.25 | 590

Now, I've improved on these practice tests over last year from between 10-50%, but I don't know if this is going to be enough to push up my score on the real Nov 2010 test since it seems I can't break 600 at home.

Any advice appreciated. I'm really at a loss on how to spend my last week of study.

---
EDIT: Corrected score of 2009 PGRE attempt.

Re: How can I break 600 on this test?

Posted: Sun Nov 07, 2010 11:11 pm
by WhoaNonstop
Honestly, no matter how bad you are at the Physics GRE, if you would have started your preparation the second you got rejected from all the other schools, you'd be sitting fairly smooth (as long as you really do understand your physics). How can you prepare with a week left? You can't really. I suggest just finding a textbook and read through it, especially on topics you are weak in. As much as you can get done "comfortably" before next Saturday. There really isn't any information that anyone can throw at you to magically get you the score you are looking for, and if there was, this information should have been discovered before this time.

-Riley

Re: How can I break 600 on this test?

Posted: Mon Nov 08, 2010 10:55 am
by vttd
Well he did say he had been studying for at least a month and a half for this particular go around.

I'm sorry to tell you (OP) that chances are not good. The fact that you've looked through the practice tests several times and have solved them by hand and still can't break the 600 barrier is not a good sign. You should do much better on the practice tests the second or third time you've taken them because you've studied those particular subjects. My best advice is to make an equation sheet and memorize as many relevant equations as possible in the last week (if you haven't already). Also make sure that you are taking the practice test under testing conditions. It looks like the most frustrating part for you is the timing, so get lots of practice working under the time constraints. Learn to let problems that you can't solve within 1 minute go.

Re: How can I break 600 on this test?

Posted: Mon Nov 08, 2010 11:12 am
by grae313
Are you making use of the solutions on grephysics.net? Look first at units or limits unless it is obvious from the question that the calculations are very quick. If you're working on a problem and it starts seeming complicated or you get stuck, mark it and skip it. Learn which questions you have a good chance of working out and which are going to be time wasters. If it seems really tough but you can eliminate a couple answer for sure, take a random guess. Your best bet to improve with only a week left is to learn how to eliminate answers without doing calculations and how to identify which questions you can answer and which you can't. Don't waste time with ones you aren't going to get right anyway. In those 25 questions you don't even see, there might be some gimmes that will boost your score.

Re: How can I break 600 on this test?

Posted: Mon Nov 08, 2010 11:48 am
by WhoaNonstop
vttd wrote:Well he did say he had been studying for at least a month and a half for this particular go around.
I realize this. Regardless of score (unless 990 of course), a month and a half of studying SHOULD increase your score.

-Riley

Re: How can I break 600 on this test?

Posted: Mon Nov 08, 2010 2:44 pm
by liutian_min
Redo the questions on the practise tests until you understand them and know how do get each question done in less than 3min. Ask some one who is good at GRE physics for help if it takes you too long on certain question.

In the real test, skip any question that looks unfamiliar or takes excessive time i.e. more than 5min.

Re: How can I break 600 on this test?

Posted: Mon Nov 08, 2010 2:46 pm
by liutian_min
One important advice:
You are just trying to do well on one particular standardized test instead of understanding all the physics behind it.
You only need to know the minimum knowledge that is enough for solving the PGRE questions.

Re: How can I break 600 on this test?

Posted: Mon Nov 08, 2010 2:49 pm
by liutian_min
For the past couple of years, it seems that the real test is easier than the practice tests.
I think you will probably do better on the easier ones since most questions are conceptual and require no calculation at all.

Re: How can I break 600 on this test?

Posted: Mon Nov 08, 2010 11:33 pm
by Blue Slime
Thanks for the suggestions. In summary:
  • WoaNonstop: Try reading textbooks on topics you're weak in
  • vttd: Make an equation sheet and memorize as many relevant questions as possible
  • vttd: Let go of any problem that you can't solve within 1 minute
  • grae313: Learn which questions are going to be time wasters
  • grae313: Learn how to eliminate answers without doing calculations
  • liutian_min: Redo practice test questions until they can each be done in >3min
  • liutian_min: Consult someone who is good at GRE physics
  • liutian_min: Skip questions that are unfamiliar or will take too much time
  • liutian_min: You only need to know the minimum knowledge that is good enough for PGRE
I've identified that the topics I could improve the most in are Thermodynamics/Statistical Mechanics and Quantum Mechanics. I'll be reading and reviewing texts on these topics.

So far I haven't been explicitly trying to memorize equations. That's something I've never did when I was doing my undergrad, but maybe for this test it is called for... I've compiled a list of highlighted equations from my practice exam study notes and will read them over daily to drill them in.

My strategy for the last practice exam, in which I skipped all questions requiring any written calculation for the first pass, seemed to prevent me from wasting all my time on a few questions during the test. However it might have been too extreme because I noticed that I didn't get around to doing some questions that only required very minor written calculations in the second pass, such as the ones dealing with relativistic momentum/energy. Need to find a balance...

Probably the most encouraging comment:
liutian_min wrote:For the past couple of years, it seems that the real test is easier than the practice tests.
I think you will probably do better on the easier ones since most questions are conceptual and require no calculation at all.
That was true of the test last year, for sure. My average score on the 4 practice exams in 2009 was 532.5, but I managed to get 590 in the Nov exam. This year my average on the 4 practice exams was 572.5, so if the proportion holds I'd expect to get a 630 on Saturday, which isn't too far from my goal.

Maybe with a focused week I have a chance. Thanks for your comments.

Re: How can I break 600 on this test?

Posted: Tue Nov 09, 2010 12:05 am
by WhoaNonstop
I urge you not to take the easier test as a reason to lack in studying. Honestly, the test is easier for EVERYONE, so it really doesn't give you a better percentile, you'll receive similar scores, regardless of how hard/easy the test is. (Unless it is easier to you and hard to everyone else).

-Riley

Re: How can I break 600 on this test?

Posted: Tue Nov 09, 2010 12:23 am
by liutian_min
Yeah, just work hard this week and it will pay off.
Do the practise tests questions and have a good freshman physics textbook at hand. Usually it contains answers to more than 60% of the questions.
I recommend Young and Freedman's University Physics because it contains all the atomic physics knowledge that is needed.

Re: How can I break 600 on this test?

Posted: Tue Nov 09, 2010 1:39 am
by CarlBrannen
The GRE is a timed test and to do well on it absolutely requires time management. One of the things I found annoying was that there was no clock on the wall at U. Washington where I took it. In addition, neither they nor I wrote down the time that the test started so the fact that I had a watch was kind of useless. I had bought a cheap watch specifically to take the test, $15 essentially wasted.

But that said, my time management was to check every 10th question with the objective of using 15 minutes. If you nail those splits you'll use 150 minutes on a first cut through 100 questions and then have 20 minutes to give everything a second look.

Re: How can I break 600 on this test?

Posted: Sat Jan 01, 2011 7:18 pm
by Blue Slime
To my absolute shock and horror, my 2010 physics GRE score was worse than the 2009 attempt. I went from a 560/22% to a 520/13%. I can't believe this result, considering:

2009
  • Studied while working full time. Couple hours a night or on the weekends only.
  • Lived abroad, no access to undergraduate textbooks, non-internet physics resources in English, etc.
  • Took test in Milan, had to endure the added stress of travel, hotel, strange location of testing center
2010
  • No job; Only focus was GRE tests and applications; studied 6-8 hours per day for a month
  • Living in hometown, nice home office, comfortable environment
  • Full access to undergraduate textbooks and all needed materials
  • Took test in hometown University, just minutes from home
I really can't believe this happened. I left the testing center feeling moderately confident. This thread was about my frustration about not reaching my modest self-set target -- but I was improving at the practice tests. I never thought I could do worse than last year, given the circumstances.

Dreams shattering... anyone who cares to see how this soap opera ends can look here.

Re: How can I break 600 on this test?

Posted: Wed Jan 05, 2011 12:29 am
by twistor
liutian_min: Redo practice test questions until they can each be done in >3min
You mean <3 min. I'd add to your list:

Learn to double check your own work, time permitting.

Re: How can I break 600 on this test?

Posted: Wed Jan 05, 2011 1:48 am
by Dreaded Anomaly
I realize any advice at this point may be too little, too late, but anyway...

The problem I see here is that you're putting down a lot of incorrect answers.
Blue Slime wrote:My statistics for this year:

Test | Answered | Correct | Raw Score | Scaled Score
GR8677 | 77 | 40 | 30.75 | 590
GR9277 | 68 | 35 | 26.75 | 580
GR9677 | 45 | 18 | 11.25 | 530
GR0177 | 52 | 33 | 28.25 | 590
You have an average rate of 51% answers correct compared to total questions answered. That needs to be a lot higher, so you need to work on catching small errors.

If you're guessing a lot, you should probably do it less; each person has an optimal guessing threshold, and not everyone's is the same. Try to keep track of which answers you would consider guesses (and why you guessed) for a practice test or section, and see how many net points they add to your score. For example, getting half of your guesses correct adds 0.75 for every two guesses (+1 for a correct guess, -0.25 for an incorrect guess).

Also, the 01 test is the most similar to the current tests, so work on being able to answer more of those questions.

Re: How can I break 600 on this test?

Posted: Wed Jan 05, 2011 2:01 am
by WhoaNonstop
twistor wrote: <3
Starting Valentine's Day early this year?

-Riley

Re: How can I break 600 on this test?

Posted: Wed Jan 05, 2011 3:19 am
by kroner
Dreaded Anomaly wrote:I realize any advice at this point may be too little, too late, but anyway...

The problem I see here is that you're putting down a lot of incorrect answers.
Blue Slime wrote:My statistics for this year:

Test | Answered | Correct | Raw Score | Scaled Score
GR8677 | 77 | 40 | 30.75 | 590
GR9277 | 68 | 35 | 26.75 | 580
GR9677 | 45 | 18 | 11.25 | 530
GR0177 | 52 | 33 | 28.25 | 590
You have an average rate of 51% answers correct compared to total questions answered. That needs to be a lot higher, so you need to work on catching small errors.

If you're guessing a lot, you should probably do it less; each person has an optimal guessing threshold, and not everyone's is the same. Try to keep track of which answers you would consider guesses (and why you guessed) for a practice test or section, and see how many net points they add to your score. For example, getting half of your guesses correct adds 0.75 for every two guesses (+1 for a correct guess, -0.25 for an incorrect guess).

Also, the 01 test is the most similar to the current tests, so work on being able to answer more of those questions.
There is never a disincentive to guess (unless you want to factor in risk aversion). If you have absolutely no idea on a question and all 5 answers are equally likely, then guessing breaks even on average. If you have any inkling whatsoever, even just some small suspicion that one of the choices seems wrong, then guessing is beneficial on average.

If you're risk averse (for example you're confident your score will be good enough and you won't have an opportunity to retake the test if it goes badly) then there is some inherent drawback to guessing you'd have to factor in. On the other hand, if you're willing to gamble for a higher score, there might be incentive to guess even on questions you haven't read.

Re: How can I break 600 on this test?

Posted: Wed Jan 05, 2011 3:36 am
by Dreaded Anomaly
kroner wrote:There is never a disincentive to guess (unless you want to factor in risk aversion). If you have absolutely no idea on a question and all 5 answers are equally likely, then guessing breaks even on average. If you have any inkling whatsoever, even just some small suspicion that one of the choices seems wrong, then guessing is beneficial on average.

If you're risk averse (for example you're confident your score will be good enough and you won't have an opportunity to retake the test if it goes badly) then there is some inherent drawback to guessing you'd have to factor in. On the other hand, if you're willing to gamble for a higher score, there might be incentive to guess even on questions you haven't read.
The fact that each question has an expectation value of zero points is true for random guessing. What I'm saying is that it might not always be a good idea to engage in non-random guessing, because one's intuition on certain types of problems or in certain situations might not be accurate.

Re: How can I break 600 on this test?

Posted: Wed Jan 05, 2011 4:02 am
by kroner
There's a big difference between your intuition being wrong sometimes and your intuition being worse than random. If by some perverse twist of fate everything you thought you knew about physics turned out to be a lie because it was always opposite day at your school and your intuitive guesses are now worse than random, this would be a good thing to discover as you take the practice tests so that on the real test you can guess the opposite of your intuition and come out ahead.

Re: How can I break 600 on this test?

Posted: Thu Jan 06, 2011 8:09 pm
by CarlBrannen
Dreaded Anomaly wrote:The fact that each question has an expectation value of zero points is true for random guessing. What I'm saying is that it might not always be a good idea to engage in non-random guessing, because one's intuition on certain types of problems or in certain situations might not be accurate.
I agree. If it were up to me, these tests would be created by having a lot of students take a "fill in the blank" test. Then I'd use the most popular incorrect answers as the incorrect multiple choices. I have no doubt that I could write a test where "random" guessing (which is really only pseudo random) would give a negative average.

Re: How can I break 600 on this test?

Posted: Thu Jan 06, 2011 9:28 pm
by kroner
Yes, it is possible to write questions where your intuitive guess is a wrong answer, and it's possible to build a test full of questions like this. However, first off the PGRE is not such a test. There are not that many trick questions.

Second (and more importantly), if there were such a test full of trick questions, once you recognized that it was built that way (assuming you have access to old tests such as you do with the PGRE) the optimal strategy would become to guess the opposite of your intuition. That strategy would be better on average than leaving questions blank. In other words, if you have information about what the test questions are like (and you do), you can always do at least as well as random, and random here is equal to leaving questions blank. Risk aversion is the only reason it might be favorable to leave questions blank.

Re: How can I break 600 on this test?

Posted: Thu Jan 06, 2011 9:59 pm
by CarlBrannen
Kroner; Some of the questions have naive answers. If I were a student about to take the practice tests, and I thought there was some possibility that I was naive, I would mark the questions I was guessing on and tote them up separately.

By the way, I once had a student get a 0 on a multiple choice test (i.e. got all questions wrong), despite answering all questions.

Re: How can I break 600 on this test?

Posted: Thu Jan 06, 2011 11:28 pm
by kroner
Of course there's always a chance you can do worse than random by guessing. That's why if you're risk averse you may be more conservative about answering questions. There's also a chance to do much better than you would on average. But the main point is that you really can't do worse than random on average if you take a reasonable approach to the test. If you think that you've naively picked the wrong answer, then pick a different answer. Your intuition about when you're being naive is testable since there are practice tests available.

I don't think I can stress enough: it's always possible to construct a guessing strategy that is more likely than not at least as good as random. Your expectation value given the information you have can't be any lower than a random guessing strategy assuming you act reasonably.

Re: How can I break 600 on this test?

Posted: Fri Jan 07, 2011 9:16 am
by WhoaNonstop
CarlBrannen wrote:By the way, I once had a student get a 0 on a multiple choice test (i.e. got all questions wrong), despite answering all questions.
I had a 10 question True-False quiz in English way back in High School and received a 0. Obviously all the questions were "tricky". =P

-Riley

Re: guessing

Posted: Fri Jan 07, 2011 2:54 pm
by Blue Slime
In reply to the discussion on guessing, I can say that while taking the practice tests with a 170 minute time constraint I marked questions as follows:

? = Guess; Not entirely convinced it is the correct answer.
L = Later; Could do it, but it will probably take too long. Come back to do it later.
S = Skip; No clue. Skip the question.

My strategy included no pure guessing or random answering, I had to either eliminate some answers or have a strong "hunch" about my unconvinced answer. I did score my guess ("?") answers separately to make sure they were improving my score, and they were. Below I'll show how the guesses added up:

Test | Answered | Correct | Guesses | Cor. Guesses | Guess % Cor. | Raw Score | Scaled Score
GR8677 | 77 | 40 | 30 | 10 | 33% | 30.75 | 590
GR9277 | 68 | 35 | 22 | 7 | 32% |26.75 | 580
GR9677 | 45 | 18 | 10 | 4 | 40% | 11.25 | 530
GR0177 | 52 | 33 | 11 | 3 | 27% | 28.25 | 590

In each case my "guessing" was better than random by a pretty good amount, leading me to believe that continuing to make these unsure answers would be beneficial to my score.

It's clear I made too many incorrect answers, but especially by the later tests they were not usually on questions I considered a guess. When reviewing the tests afterwards, there were plenty of "oh my f***ing god" moments when I realized I had made some trivial or embarrassing mistake. I wish I knew how I could systematically eliminate those kinds of errors. It's always the kind of thing that I can catch if I have more time, but for the PGRE I always have to go with my first assumptions.
twistor wrote: Learn to double check your own work, time permitting.
I wish! I never had time left over at the end of the test. Usually by about question 50 I'd realize I was not on pace to even consider every question, and it'd be a sprint to speed-read till the end, hoping for some easy ones. In some of the practice tests I had 10 minutes or so to try a couple of "Later" questions, but I think in the real test it was down to the wire.

Re: How can I break 600 on this test?

Posted: Fri Jan 07, 2011 3:33 pm
by grae313
kroner wrote: However, first off the PGRE is not such a test. There are not that many trick questions.
Carl is right, it's not that there are trick questions, but naive answers and answers that one would get by making common mistakes. I remember on the exam leaving out a subtlety in a derivation and arriving at an answer that was one of the multiple choice options. I selected it, and then later on in the test another problem jolted my memory and I realized I had left something out, redid the problem and arrived at (hopefully) the correct answer. I also did analysis on my practice exams to determine how well I did by guessing. Even when I was able to eliminate one or two answer, I found that I very often guessed wrong when I chose the answer that "looked good" and that the guessing after eliminating a few options did not always help my score.

My advice would be to eliminate the answer you know to be wrong and then guess among the remaining without looking at the options. Just choose a random letter. There are misleading-looking answers out there.

Re: How can I break 600 on this test?

Posted: Fri Jan 07, 2011 8:01 pm
by CarlBrannen
WhoaNonstop wrote:I had a 10 question True-False quiz in English way back in High School and received a 0. Obviously all the questions were "tricky". =P -Riley
Wow! And did you make your score known to the whole class like my student did?

By the way, I recall another thing the same guy did. Since I'm considerably older than most of my students I sometimes give them helpful advice. The classes at ITT are >3 hours long, once per week, and you have to break the class up with breaks and stuff like this in order to prevent vapor lock in their brains.

So in the "great advice from Carl" minute, I mentioned that an excellent place to meet possible dating partners was at the county's free VD clinic. (This is for a bunch of reasons you can fill in for yourself, probably.) He piped up and mentioned that he did meet a girl there. And about 30 seconds later added that he was only there for a check-up.

It's nice to watch your students graduate so I attend the ceremonies (which are quarterly at ITT). I got to shake his hand at graduation this past quarter. And yes, after shaking the hands of lots of people I do have a habit of giving my hands a good washing.

Also, I wonder if there's a correlation between the utility of guessing and how high your score was. My belief is that with myself, guessing was not a good idea. But I do it anyway. On the other hand, I feel a great motivation to fill in all the answers as it's the only way to get 100%.