What to do with the minutes left at the end of a module: the three-pass review system, Reading and Writing trap tactics, the four-step math answer check, how to budget review time, and the evidence rule that decides when to change an answer and when to keep it.
Subject: SAT Prep · 76 slides · applied lesson
Open the interactive version of this deck · Homework for this lesson
Title
SAT Prep · Review Strategy
What to do with the minutes left at the end of a module
Objectives
This deck is about the last five minutes of a module — the most under-used time on the whole test.
A few minutes of review can mean a few extra points — but only if the minutes are spent in the right order.
Section
Section 1
Warm-up
Answer honestly — the answer determines how much this deck is worth to you.
Discussion prompt
Last time you finished a module with time to spare, what did you actually do with it? Be specific about the first thing you did.
Answer:
The three common answers, in order of frequency: sat still, started at question 1 and re-read everything, and stared at the one question that had bothered you.
All three are worth roughly zero. The first spends nothing, the second spends everything on questions that were already right, and the third spends everything on your least likely point.
Concept
There are two reasons to get a question wrong: you did not know how, or you knew how and slipped.
Studying fixes the first. Only review fixes the second — and on a timed test, the second is a large share of the misses.
careless error — A miss on a question you could have answered correctly with no additional knowledge. The distinguishing test: when it is explained, you learn nothing.
Estimation
Go back through your last practice test in your head — the questions you got wrong but understood immediately when you saw the answer.
Predict first
Roughly what fraction of your wrong answers were careless rather than knowledge gaps?
Correct: For most students in timed conditions it is around a third — treat that as a working assumption and then measure your own.
Why: The exact fraction varies by student and it is not worth arguing about; what matters is that it is large enough that reviewing beats studying for those particular points. Your own number is measurable: after your next practice test, tag each miss as 'did not know' or 'knew it, slipped'. Do that twice and you will have a personal figure far more useful than any general one — and it tells you how much of your remaining score is sitting in the last five minutes of each module.
Concept
Notice the split: finding is about work already done, preventing is about work not finished. They need different passes, which is why there are three.
Real world
You have a functioning review process somewhere in your life. Let us borrow it.
Discussion prompt
Think of something you already check before submitting — a text before you send it, a lock before you leave, a bag before a trip. What makes that check work, and how long does it take?
Answer:
Good checks share two properties: they are specific (you check named things, not a general vibe), and they are short (a list you can finish, not an open-ended worry).
A bad review is the opposite — vague and unbounded. 'Let me look over my answers' has no stopping condition, so it either stops at random or never stops. Every pass in this deck has both a target and a time limit.
Two truths and a lie
Four things students believe about the end of a module. Two hold up.
Sort into buckets
Sort each claim.
Section
Section 2
Concept
Use this whenever you finish a module with time remaining. The order is the whole design.
Cheapest and most certain first. Pass 1 recovers guaranteed zeros; pass 2 gambles minutes on maybes.
Concept
Open the Review page. You are not solving anything in this pass.
Any blank you find here is a guaranteed zero converted into a chance for the cost of one tap. Nothing else in the review is that cheap.
Concept
Now you work. Flagged questions and ones you felt uneasy about — nothing else.
Stem first, work second. If you solved the wrong question, checking your arithmetic will confirm the wrong answer perfectly.
Concept
Mechanical, fast, and the one students skip.
On the digital test this replaces the old bubble-sheet check. The failure mode moved, it did not disappear: an answer you typed but never confirmed scores exactly like an empty bubble.
Ranking
You have six minutes left. Put the passes in the order that recovers the most points.
Put in order
Why: The scan comes first because it is the only pass with a guaranteed return: filling a blank converts a certain zero into a real chance in about two seconds. Deep review comes second because it is where the uncertain minutes belong, and it needs the triage the scan produced. The final check is last by necessity — it has to verify the answers the deep review just changed, so running it earlier would check a state that no longer exists.
Estimation
You finish a module with exactly 6 minutes on the clock and 4 flagged questions.
Predict first
How should those 6 minutes be split across the three passes?
Correct: Roughly 1 / 4 / 1 — a fast scan, the bulk on the flags, one minute to confirm.
Why: With only four flags the scan is quick: you are checking for blanks and reading a short list, which is well under two minutes. That buys the deep review four minutes, or about a minute per flag, which is realistic. The final minute is not optional — it is what catches an answer you changed during pass 2 and never confirmed. The one split that is always wrong is 0 / 6 / 0: skipping the scan means you may spend all six minutes on a flag while a blank sits somewhere else in the module.
Trap
Five minutes left. You go back to the beginning.
Re-read question 1, which you answered confidently in 30 seconds
Why: It was right the first time and it is right now. The check produced no information.
Work forward through questions 2 to 6
Why: All confident, all correct, all re-verified at a cost of about four minutes.
Time is called at question 7
Why: The three flagged questions at 11, 14 and 19 were never reached.
Five minutes spent, zero points recovered.
Five minutes left. You open the Review page.
Scan for blanks — find one at question 20, and fill it
Why: Ten seconds spent, one guaranteed zero converted into a real chance.
Go straight to the flags: 11, 14, 19
Why: These are the only questions where your answer might change, so they are the only ones worth minutes.
Finish with a one-minute confirmation pass
Why: Every answer, including the two you changed, is registered.
Five minutes spent, two to three points in play.
Faded example
Fill the blanks from memory.
Fill in the blanks
Pass 1, the Quick Scan, takes 1-2 minutes and looks for blanks and flags. Pass 2, the Deep Review, takes 3-5 minutes and re-reads the question stem before checking the work. Pass 3, the Final Check, takes 1 minute and confirms every answer actually registered.
Why: Each pass is defined by a target and a time limit, which is what stops review from becoming an anxious wander through the module. Notice that the stem re-read sits in pass 2 rather than pass 1: the scan is deliberately not a reading pass, because reading twenty-two stems is exactly the thing that eats five minutes and returns nothing.
Section
Section 3
Concept
| tactic | what it catches |
|---|---|
| Re-read the question stem | Misreading — cover your chosen answer first so it cannot bias the reading |
| Return to the passage | Answers that are plausible but not actually supported by the text |
| Eliminate trap answers | Choices that are too extreme, too narrow, or out of scope |
| Check vocabulary in context | Words that fit the meaning but not the tone of that specific sentence |
Every one of these is a return to the source. Reading answers on their own is how wrong ones start to sound reasonable.
Intuition
The most common Reading error is choosing an answer that is true — just not something the passage said.
The test for every Reading answer is physical: can you put a finger on the words that make it right?
If the justification is 'well, it makes sense', you are reasoning about the world. The question is about the text.
Discrimination
The passage describes a scientist who repeated an experiment eleven times before publishing.
Sort into buckets
Which inferences are supported by that sentence alone?
Concept
| trap | how it works |
|---|---|
| The 'almost right' answer | A near-correct distractor placed to catch rushed readers |
| Extreme language | always, never, only, must — rarely survive contact with a passage |
| Mismatched tone | Factually accurate but too formal, too casual, or emotionally off |
| Out-of-scope inference | Requires assuming something the passage does not say |
Three of the four are detectable without re-reading the passage — which makes them the fastest checks in a deep review.
Elimination
The passage argues that a policy had mixed results: measurable benefits in cities and no detectable effect in rural areas.
Eliminate the wrong options
Which choice best describes the author's view of the policy?
Survives elimination: b
Why: The surviving choice is the boring one — it restates exactly what the passage said and adds nothing. That is almost always the shape of a correct Reading answer, and it is why students talk themselves out of it: the right answer often feels too obvious to be worth a point. Note that three of these four were eliminated purely on the language of the choice, without going back to the passage at all.
Trap
You return to a flagged Reading question in your deep review.
Re-read the four answer choices first, to remind yourself of the question
Why: Now all four are loaded in your head, and each one is a hypothesis looking for evidence.
Skim the passage looking for support
Why: You find something that resembles choice C, because a long passage contains something resembling almost anything.
Change your answer to C
Why: This is confirmation bias with a timer on it — you searched for a match rather than reading for meaning.
You return to a flagged Reading question in your deep review.
Cover the choices. Re-read the question stem alone.
Why: The stem is the only thing that defines what a right answer must do, and it is short.
Re-read the relevant lines and answer in your own words first
Why: Your own words are unbiased by the distractors, and they give you something to match against.
Now uncover the choices and find the one that matches
Why: Matching a formed answer to a list is fast and resistant to traps — the reverse direction is neither.
Explain it to yourself
The single most repeated instruction in this section is 'cover your chosen answer'.
Discussion prompt
Why does covering the answer you already picked make re-reading the stem more effective? What exactly goes wrong if you leave it visible?
Answer:
A visible answer turns re-reading into verification instead of reading. You read the stem looking for reasons your choice was right, and a stem will nearly always supply one if that is what you are hunting for.
Covered, you read the stem to find out what it asks. Those are genuinely different mental operations, and only the second one can find a misreading.
Matching
Tactics are not interchangeable — each one is tuned for a particular failure.
Match the pairs
Why: The first three are checks on the choice; the fourth is a check on the question. That distinction is the reason the stem re-read comes first in a deep review: if you answered the wrong question, no amount of scrutiny applied to the choices will reveal it, because one of them really is the best answer to the question you thought you were reading.
Section
Section 4
Concept
Plugging back in is the strongest of the three: it is a proof, not a re-check. Re-doing your work only reproduces your mistakes.
Worked example
The question: solve for x. Your answer was x = 4.
\[ \frac{2x + 6}{x - 1} = 5 \]
Do not re-do the algebra — substitute 4 into the original equation
Why: Re-doing the work tends to repeat the original slip, because you follow the same path. Substitution tests the answer independently of how you got it.
\[ \frac{2(4) + 6}{4 - 1} = \frac{14}{3} \approx 4.67 \neq 5 \]
The check fails, so x = 4 is wrong — now solve it properly
Why: Multiply both sides by (x - 1): 2x + 6 = 5(x - 1), so 2x + 6 = 5x - 5, giving 11 = 3x.
\[ x = \frac{11}{3} \]
Verify: substitute 11/3 back into the original
Why: The numerator is 2(11/3) + 6 = 22/3 + 18/3 = 40/3, and the denominator is 11/3 - 3/3 = 8/3. Dividing gives 40/8 = 5, which matches the right-hand side exactly.
The whole check took about twenty seconds and it turned a wrong answer into a right one.
Prediction
Substitution is powerful but it is not universal.
Predict first
You substitute your answer into the original equation and it works perfectly. What kind of error could still be hiding?
Correct: Answering a different question — substitution verifies the equation, not the request.
Why: If the question asked for the value of 2x and you found x, your x will satisfy the equation beautifully while your answer is still wrong. Substitution is a check on your algebra, not on your reading, which is exactly why the stem re-read is a separate step and comes first. The other two options are precisely what substitution does catch: any arithmetic or sign error in your solving shows up as a failed check.
Concept
Step 1 catches the most points and takes the least time. Students start at step 3.
Ranking
These are the four checks, shuffled.
Put in order
Why: The order runs from most expensive error to least. A misread question invalidates everything downstream, so it is checked first and costs three seconds. A wrong setup invalidates all the arithmetic built on it. Only once the question and the structure are confirmed is there any point in checking the computation — and the answer form is last because it only applies to an answer you have already decided is correct.
Error analysis
A student solved this, checked it, and still got it wrong. The question asked for the value of 3n.
Annotate
On: \( 5n - 8 = 27 \;\Rightarrow\; 5n = 35 \;\Rightarrow\; n = 7 \;\Rightarrow\; \text{answer: } 7 \)
Trap
You return to a flagged math question with 90 seconds.
Re-solve the problem from the beginning
Why: You take the same path, make the same reading, and reach the same place.
Get the same answer and feel reassured
Why: Agreement between two runs of the same flawed process is not evidence — it is repetition.
Move on with the wrong answer, now more confident
Why: Worse than not checking, because the false confidence removes the question from your next review.
You return to a flagged math question with 90 seconds.
Re-read the stem, then substitute your answer into the original
Why: Substitution is independent of how you solved it, so it can disagree with you.
The check fails — now you know there is an error and roughly where
Why: A failed check localises the problem: the setup or the arithmetic, and the setup is quicker to inspect.
Fix and re-verify
Why: The second verification tests the corrected answer, closing the loop.
Comparison
Two cells are blank. Complete them from what you just learned.
Comparison matrix
| step | what it catches | time |
|---|---|---|
| 1. Re-read the question | Solving for the wrong quantity | 3 seconds |
| 2. Check the setup | A wrong equation or relationship | 10 seconds |
| 3. Verify arithmetic | Sign errors, slips, order of operations | 20 seconds |
| 4. Confirm the form | Unsimplified or unroundable answers | 5 seconds |
Missing information
A student's work reads, in full: 'x = 12'.
Discussion prompt
You are asked to review it. What do you need before you can check anything, and what does that tell you about how to write work during the test?
Answer:
You need the original equation and what the question asked for. Without them, 'x = 12' cannot be checked by anyone, including the student who wrote it.
The practical lesson: your scratch work is written for the version of you who will return in four minutes. Write the setup line, not just the result — that is what makes the review pass fast rather than a fresh start.
Section
Section 5
Concept
How you spend the leftover time determines how many points you actually recover.
| what | share of review time | why |
|---|---|---|
| Flagged and skipped questions | about 55% | The only questions where your answer is genuinely undetermined |
| Answers you felt uncertain about | about 25% | Undetermined, but you at least reached something |
| Confirming answers registered | about 12% | Cheap, mechanical, catches guaranteed zeros |
| Answers you were confident about | about 8% | Nearly always right — a token pass at most |
Read the last row again. Eight percent, not zero — but nowhere near the half of your time it usually gets.
Trade off
Fill in the blanks: what does a minute spent on each category actually return?
Comparison matrix
| spend a minute on | expected return | risk |
|---|---|---|
| A flagged question you half-solved | High — the setup already exists | Low |
| A blank you never started | Low for the answer, but the guess itself is free | Low |
| A question you answered confidently | Very low — it is probably already right | High — this is where correct answers get changed to wrong ones |
| Confirming answers registered | Moderate and certain | None |
Concept
The goal is efficient recovery, not perfection. You are not trying to be certain, you are trying to convert.
Elimination
Three minutes left. One blank at Q20, two flags at Q6 and Q17, and a nagging feeling about Q3 which you answered confidently.
Eliminate the wrong options
What do you do first?
Survives elimination: b
Why: Filling the blank takes about three seconds and converts a certain zero into a real chance — nothing else in the review has that return per second. Only then do the flags get the remaining time, and the confident question with a vague bad feeling gets nothing at all. Ordering by cost-to-fix rather than by question number or difficulty is the whole skill.
Trap
Four minutes left, three flags. You open the first one.
Work it for 90 seconds, then think 'I'm close'
Why: The same sunk-cost feeling from the 60-second rule, now with less time to absorb it.
Spend another 90 seconds on it
Why: Three of your four minutes are now inside one question, which you still have not answered.
Time is called with two flags never opened
Why: Two questions that might each have taken 40 seconds were traded for one that took three minutes and failed.
Four minutes left, three flags. You open the first one.
Give it 90 seconds, hard cap
Why: The cap is set before you start, when setting it is free — the same reason pre-deciding works.
At 90 seconds, put down your best answer and move to the next flag
Why: The answer is banked. If time remains after the third flag, you can come back.
All three flags get opened
Why: Three chances instead of one, and the two cheap ones were probably the two you could actually finish.
Edge cases
The three-pass system assumes five or six minutes. It will not always be there.
Discussion prompt
You finish a module with exactly 60 seconds left. Which pass survives, and what do you drop entirely?
Answer:
Keep the scan. Sixty seconds is enough to open the Review page, find any blank, and fill it. That is the highest-return action on the list and it is nearly free.
Drop the deep review. With one minute you cannot re-read a stem, re-work a problem, and reach a considered decision — you can only rush one, and a rushed change is more likely to hurt than help.
The general principle: when time collapses, do the certain thing, not the valuable thing. Filling a blank is certain; changing an answer is a gamble.
Section
Section 6
Concept
Changing answers can help or hurt. Which one it does depends entirely on why you are changing.
the golden rule — Change only with evidence. Intuition alone is not a reason to switch — if you cannot name the specific error, keep what you have.
This is the highest-stakes rule in the deck, because it is the only one that can actively lose you points you already had.
Concept
Every one of them is nameable. You can say the sentence out loud: 'I misread the units', 'choice B says always'.
Concept
The last one is worth its own note: a second change is almost always thrashing, not reasoning.
Sorting
Six situations from a real review pass. Sort each one.
Sort into buckets
Does this justify changing the answer?
Commit first
You are reviewing a Reading question. Your answer is C. On second look, A seems more elegant, but you cannot find anything wrong with C.
Predict first
What do you do?
Correct: Keep C — you have found no error, and elegance is not evidence.
Why: The only justification for a change is a named defect in the answer you have, and 'A seems more elegant' is a statement about A rather than a criticism of C. Changing back and forth is the thrashing pattern that wastes the whole review pass. And blanking is strictly worse than any answer, since there is no wrong-answer penalty. Note how confident you felt: students who feel very confident here and still change are the ones who lose points to review rather than gain them.
Counterexample
Rules you cannot challenge do not survive contact with a real test.
Discussion prompt
Construct the strongest case you can for changing an answer without a nameable error. When, if ever, is unease actually informative?
Answer:
The strongest case: unease sometimes is a half-formed recognition — you noticed something and have not yet articulated it. That is real.
But the response is not to change the answer, it is to spend fifteen seconds trying to name it. If the name arrives, you now have evidence and the rule permits the change. If it does not arrive, the unease was noise.
So the rule survives, in a sharper form: unease is a reason to look, never a reason to switch.
Section
Section 7
Concept
Five items, roughly forty seconds. Run it once, at the end of the module.
Concept
On the digital SAT, Desmos is available on both math modules — so 'verify the arithmetic' costs almost nothing.
Matching
Every checklist item exists because of a specific, common miss.
Match the pairs
Why: None of these four errors is a knowledge gap — every one of them happens to students who fully understand the underlying mathematics. That is precisely why a checklist works here and would not work for a topic you have not learned: a checklist can only catch errors you are capable of noticing once your attention is pointed at them.
Notation
Half of all 'careless' math misses are a misreading of this one line.
Annotate
On: \( \text{If } 2x + y = 14 \text{ and } y = 4, \text{ what is the value of } 3x? \)
Definition probe
Checklists only help if you run the right one.
Sort into buckets
Sort each review action.
Section
Section 8
Concept
Time spread evenly over twenty-two questions is time not spent on the four that were undetermined.
the fix — Triage first — only revisit flagged or uncertain answers. Confident answers get a glance at most, and only if minutes remain.
Concept
Under stress, students change right answers to wrong ones. Anxiety is not evidence of an error.
the fix — Require a specific, articulable reason before changing. If you cannot say the sentence out loud, do not touch the answer.
Concept
An answer you selected but never confirmed scores exactly like a blank. This survived the move to digital — it just changed shape.
the fix — Always end with the one-minute pass over the Review page. It is the highest return-on-investment minute in the module.
Concept
Quickly scanning answers is not reviewing them. A pass that reads nothing catches nothing.
the fix — Read every flagged question stem fully before checking any work. Fewer questions, read properly, beats all of them read badly.
Concept
Students run out of review time because they start at question 1 and never reach the flags.
the fix — Open the Review page first. Go to flags before anything else — that is what you flagged them for.
Concept
The SAT has no wrong-answer penalty. A blank is a guaranteed zero; a guess is a real chance.
the fix — Fill in every answer before time is called — even a pure guess strictly beats a blank.
Ranking
All six are worth fixing. Only one order reflects the damage.
Put in order
Why: Panic-changing ranks worst because it is the only mistake that actively subtracts points you had already earned — every other item on this list is a failure to gain. Blanks come second: certain zeros, but at least they were never points. Reviewing everything and rushing are both wasted-opportunity errors, and rushing is last because a rushed review at least occasionally stumbles onto something.
Comparison
Two fixes are missing. Fill them in.
Comparison matrix
| mistake | the fix |
|---|---|
| Reviewing every question | Triage first — flagged and uncertain only |
| Panic-changing correct answers | Require a specific, articulable reason before changing |
| Skipping the confirmation pass | Always end with the one-minute Review page pass |
| Leaving blanks | Fill in every answer — a guess strictly beats a blank |
Explain it
One of the six rules reliably gets argued with.
Discussion prompt
A friend says: 'If I think an answer is wrong, I should change it — trusting my gut has worked for me before.' Give the response, and be fair to the part they have right.
Answer:
Fair part: a gut feeling is sometimes a real recognition that has not surfaced yet. It is worth fifteen seconds of investigation.
The response: the change should follow the investigation, not the feeling. If fifteen seconds produces a nameable error, change it — you now have evidence. If it does not, the original answer was chosen with more time and less pressure, and it is the better bet.
The rule is not 'never change'. It is 'change on evidence, not on feeling' — and that distinction is the entire content of the rule.
Section
Section 9
Pattern
This is the reusable recipe. Run it at the end of every module, in every practice test, until it is automatic.
Eight steps, two of which are about stopping. That is the same shape as the 60-second rule, and it is not a coincidence.
Check
Answer from the prioritisation rules, not from instinct.
Check your understanding
You finish a module with 5 minutes left. You have 1 blank, 3 flagged questions, and 15 questions you answered confidently. What is the correct first move?
Answer: A
Why: Filling the blank takes seconds and converts a guaranteed zero into a real chance, which is the highest return per second available. The flagged questions come next because they are the only other place where your answer is genuinely undetermined.
Check
The golden rule, applied.
Check your understanding
During review you re-read a math question and realise it asked for the perimeter, but you calculated the area. What should you do, and why?
Answer: A
Why: This is exactly the case the golden rule is built for: you can name the defect in one sentence — 'I solved for area, the question asked for perimeter.' A nameable error is evidence, and evidence is the one thing that justifies a change.
Check
One question about the pass students skip.
Check your understanding
Why does the final one-minute pass still matter on the digital SAT, given that there are no bubble sheets to misalign?
Answer: A
Why: The failure mode moved rather than disappearing: a grid-in you started typing and abandoned, or an answer changed mid-edit when you navigated away, registers as unanswered. The Review page shows the true state of every question, which is what the final pass is checking against.
Hypothesis
Commit to a number before your next practice test. This is how you find out whether the system is working.
Predict first
On your next full practice test, how many raw points do you predict the three-pass review will recover across all four modules?
Correct: Most students recover 2 to 4 across a full test — but the prediction matters more than the number.
Why: Write the prediction down and then actually measure it: mark every answer you change or fill during a review pass, and afterwards count how many of those were right. That count is the only honest measure of whether your review is helping or hurting. If more than a couple of your changes went from right to wrong, the problem is not your review speed — it is that you are changing on feeling rather than on evidence.
Connect it up
From memory, and without scrolling back.
Draw it
Sketch the three-pass review as a timeline across the last five minutes of a module. Mark where the blanks get filled, where the flags get worked, where the 90-second cap fires, and where the confirmation pass sits. Add the one-line golden rule for changing an answer.
Exit ticket
One last commitment before you close the deck.
Predict first
Which part of the review process are you most likely to skip under pressure?
Correct: Whichever you picked is the one to rehearse deliberately in your next timed practice.
Why: There is no wrong answer here — the value is in naming your own failure mode before test day, because a named habit is one you can catch yourself doing. If you picked the golden rule, add a physical step: say the error out loud before your hand moves. If you picked pass 3, set the habit of opening the Review page the instant you answer the last question, so the confirmation pass has somewhere to start.
Recap
Six lines. Practise them on every timed mock until they are automatic.
| you can now | the move |
|---|---|
| Spend leftover minutes well | Scan, then flags, then confirm |
| Catch a misread question | Cover your answer, re-read the stem |
| Verify a math answer | Substitute back into the original |
| Kill a Reading trap | Extreme, out-of-scope, or tone-mismatched |
| Decide on a change | Name the error out loud, or keep what you have |
By test day this should feel automatic — which only happens if you run it on practice tests, not just on the real one.
Want this taught 1-on-1? Alexander tutors SAT Prep — $55/session, free consultation.