Why Your Practice Scores Bounce Around (And Why That's OKAY)

Say you took a full practice GRE on Saturday morning. Last week you scored 160 in Quant. This week you scored 158. You spent the rest of the weekend convinced that you are getting WORSE, that your study plan is broken, and that maybe you are just not a "math person."

Stop. You almost certainly did not get worse. A 2 point drop on one practice test is not news. It is noise. And if you react to noise like it is news, you will make bad decisions about how to study, which is how people actually DO get worse.

Here is why practice scores bounce around, and what you should be looking at instead.

A practice test is a measuring device, not a verdict

A practice test gives you an estimate of where you are right now. It does NOT make you a better Test Taker, and one result on its own tells you very little. Think of it like stepping on a scale at the gym. Your weight moves a little from day to day based on water, sleep and what you ate. Nobody who trains seriously panics over one morning's number. They watch the trend over weeks.

Your GRE scores work the same way, and there are a few clear reasons why.

1. The scale is tight

Your Quant score and your Verbal score each come from two scored sections of 20 questions, so 40 questions per subject. Those 40 questions get turned into a scaled score, which is the 130 to 170 number you see at the end. That is a 41 point range built from 40 questions. Do the math on that for a second. A couple of questions going the other way can move your scaled score a point or two. Two silly mistakes is all it takes to turn a 160 into a 158.

2. The test adapts section by section

The GRE is adaptive section by section. That means how well you do on your first Quant section decides how hard your second Quant section will be. A rough first section sends you to an easier second section, and an easier second section caps how high you can score. So a couple of sloppy errors early on do not just cost you those questions. They change the whole second half of that subject.

3. Your conditions changed

Did you take both tests the same way? Same time of day, same amount of sleep, both essays written, the break taken at the same point? If you skipped the writing on one test, or did it at night after work, you did not run the same experiment twice. The real GRE takes about 3 hours and 45 minutes, and if you are not practicing for that full stretch, read why you have to train for the FULL 3 hours and 45 minutes.

4. You saw different questions

One test might hit you with three probability questions. The next might have none. If probability is a hole in your knowledge, your score will swing based on luck of the draw, not on your actual ability.

Now here is the wager. I'll bet that if you pull up that 158 test, you can find the 2 or 3 questions that cost you those points. And I'll bet MOST of them were questions you knew how to do.

Watch how one assumption costs a point

Let's prove how easy it is to drop a question you could have gotten. This is a Quantitative Comparison question, a GRE Quant type where you compare two amounts and pick which is bigger, or decide that you cannot tell.

Notice the irony. That question is about averages hiding a lot of variety. Your own practice scores do the same thing.

What to look at INSTEAD of the score

The number at the end is the least useful thing a practice test gives you. The useful stuff is question by question. After every practice test, go through EVERY question, right and wrong, and write down three things:

• Was it gettable? Did you know the concept and still miss it?
• How long did it take? A correct answer that took 4 minutes is still a problem.
• What does your scratch paper show? Did you take notes, or did you do it in your head?

Do that for a few tests and patterns jump out. THAT is real data. A single scaled score is not.

The mistakes I see people make after a "bad" score

1. Treating one result like a trend. You drop 2 points and rewrite your whole plan. INSTEAD: Compare averages across several tests taken weeks apart. Small moves up and down mean you are performing at the same general level.
2. Taking MORE practice tests to "fix" it. Three tests in one week will not make you better. It will make you tired and give you more noise. INSTEAD: One full practice test per week at most. Spend the days in between training on what the review showed you.
3. Only reviewing the questions you got wrong. Some of your right answers were lucky guesses or took way too long. INSTEAD: Review every question, and flag anything you guessed on or that ate too much time.
4. Blaming the "hard" questions. It feels better to say the test got harder. INSTEAD: Count how many points you lost on questions you KNEW how to do. I'll bet that number is bigger than the number you lost on truly hard ones.
5. Running tests under fake conditions. Skipping the essays, pausing for a snack, taking it in two sittings. INSTEAD: Take it like Test Day, every time, so the results are actually comparable. If your weekdays are packed, build that into your week the way we describe in a full time job is NOT an excuse, it's a schedule.

So when SHOULD you worry?

Here is a forced choice. What do you want MORE: to feel good about this week's number, or to know exactly what to fix? You cannot have both from one test.

Worry when the same problem shows up test after test. Worry when your scratch paper is nearly empty. Worry when your scores stay flat for weeks while you keep studying the same way. Those are real signals. A 2 point dip with a clean, specific reason behind it is not a signal. It is a to-do list.

And remember what the practice tests are for. ETS gives you free POWERPREP practice tests that are built like the real thing, which makes them great measuring devices. When Test Day comes, the GRE will give you the score you EARN, and you can see how scores and ScoreSelect work on ETS's scores page. The way you earn it is by training on the specific causes behind your missed points, not by staring at a number.

Your move, today

Pull up your last two practice tests. Don't read another word of anything else until you do. Sort every question you missed into three piles: silly mistake, knew it but too slow, and truly did not know it. Count each pile. Then pick the BIGGEST pile and spend your next study session on it, and only it.

I'll bet the "silly mistake" pile wins. If it does, you just found your next 2 points, and they were never really gone.