Choosing between algebra, graphing, and mental calculation
Reading position saved on this browser; this is not a completion record.
Start here
Choose the shortest reliable path, not the most impressive tool.
By the end of this section
Recognize when structure removes computation; use hybrid methods for harder tasks; manage a timed module deliberately; and respond appropriately to mathematical errors versus testing-device problems.
5.1 A method-selection decision table
Your goal is an answer supported by correct reasoning. A method’s real cost includes setup, calculation, interpretation, and checking. A graph that takes five entries and an awkward window may be slower than one line of algebra. A calculator may be faster when the remaining work is merely a long decimal calculation.
What you notice | Good first method | Why |
|---|---|---|
Friendly arithmetic or a direct ratio | Mental calculation | Avoids unnecessary entry and transfer. |
A requested expression already grouped in an equation | Algebraic structure | May avoid solving for the variable. |
Two awkward equations with numerical coefficients | Graphing or elimination | Intersections can reduce arithmetic, but label the coordinates. |
An exact number of roots or a parameter condition | Algebra, then graph | A display can suggest a result; exact reasoning establishes cases. |
Repeated evaluation at several inputs | Function and table | One definition serves all inputs. |
A stated model fitted to several observations | Regression | Appropriate when the data and model identify the desired parameters. |
A long power or a non-special angle | Numerical calculation | Performs the arithmetic after the mathematical setup. |
A causal or sampling claim | Study-design reasoning | A fitted coefficient cannot repair a flawed design. |
Practice a decision before an entry
Ask, “What quantity is requested, and what would make it visible?” For a root, look for y
Speed is personal; reliability is nonnegotiable
Compare methods on problems you have not memorized. Record errors as well as time. A method that feels fast but produces wrong-coordinate or domain mistakes needs repair before it becomes your default.
5.2 Hybrid solutions: structure first, computation second
A strong calculator solution often begins on scratch paper. Define the variable, write the relationship, identify restrictions, and then let the tool handle the part it performs well. Finish with a mathematical check.
Three productive combinations
Simplify, then evaluate. For the area of a circle with radius , simplify to 50π before calculating. This preserves precision and reduces entry length.
Graph, then certify. An intersection near (5, 4) is a candidate. Substituting both coordinates into both original equations certifies it when those exact values work.
Locate a threshold, then test neighboring integers. A continuous graph may cross a target between days 11 and 12. A question about whole days requires testing day 11 and day 12, not reporting the graph’s decimal crossing.
Original explanatory graph. The exact evaluations, not the drawing, decide whether a strict threshold is met.
Do not let a useful shortcut erase a case
For ax2 + 6x + 9
A method is complete only when it answers the original question
Exact or approximate? One solution or all solutions? Input or output? Continuous value or integer? Model prediction or observed measurement? State the distinction before you finalize the answer.
5.3 Managing a Bluebook Math module
Under standard timing, SAT Math contains 44 questions in two 35-minute modules. Each module contains 22 questions. That is about 95 seconds per question on average, not a recommended time limit for every question. Approved timing accommodations change your schedule; rehearse with your own settings.[15]
You can revisit questions within the active module, but you cannot return to a module after leaving it. Use the question menu to distinguish unanswered items from questions merely marked for review. Marking is an organizational action, not a response.[2][20]
A flexible three-pass routine
Pass | Purpose |
|---|---|
First: collect | Solve questions with a clear path. If a question stalls, preserve useful work, choose your best current response when possible, mark it, and continue. |
Second: resolve | Return to unfinished questions where you have a promising next step. Avoid repeating the same failed graph setup without changing your approach. |
Third: audit | Check unanswered items, response transfer, requested quantities, and specific suspected mistakes. Do not replace a justified answer merely because it feels unfamiliar. |
This is a practice framework, not an official prescription or a guarantee of a higher score. Calibrate it using your own full-length practice. Some students work most accurately in order; others benefit from an early return pass.
Use the timer without letting it consume attention
Bluebook lets you hide the timer until the final five-minute alert.[2] Establish a few planned checks rather than looking after every arithmetic step. A useful final check is whether every intended answer is actually selected or typed.
Do not try to reverse-engineer the adaptive route
The second module is tailored using first-module performance, and scoring uses performance across both modules. College Board advises trying your best on every question. Do not deliberately miss items, assume a difficult-looking question is unscored, or spend working time guessing your eventual score.[21]
Protect the module boundary
A marked question does not follow you into the next module. Before time expires, use the question menu and review screen to confirm that blanks and flagged items have received deliberate attention.
5.4 Troubleshooting without abandoning the mathematics
First identify the kind of problem. An incorrect calculator expression, an unhelpful graph window, and a testing-device failure require different responses.
What happened? | Appropriate next step |
|---|---|
An expression reports an error | Inspect parentheses, denominators, roots, variable definitions, and function spelling. A syntax error is not proof of no solution. |
A graph looks empty | Confirm the expression is enabled, then check axis bounds and the expected scale. Verify that the plotted relationship is real-valued on the chosen interval. |
A coefficient appears wrong | Check the table pairs, column names, model form, and whether an old definition fixed a parameter you intended to fit. |
An output seems implausible | Check units, signs, angle mode, and the original target before re-entering the same expression. |
Bluebook or the device is unresponsive | Raise your hand and follow the proctor’s and app’s troubleshooting instructions. Do not improvise with a phone, external browser, or unauthorized device.[20] |
Separate mathematical recovery from technical recovery
When only your graphing approach is failing, a short algebraic method may keep the work moving. When the testing app or device fails, alert the proctor. Do not assume the timer paused, your work disappeared, or you are authorized to restart or change devices independently.
Rehearse the full transfer chain
Use a preview to locate controls. Then use an official full-length practice test to rehearse the cycle: read, model, calculate, check, enter, navigate, and review. The preview itself is unscored; scored practice is a separate option.[16]
The durable five-step routine
Target: name the quantity. Model: write the relationship and restrictions. Tool: choose and operate a method. Verify: test meaning and accuracy. Transfer: enter the answer in the question, not only in the calculator.
15 worked examples
5.01. Let friendly percentages stay friendly
What is 18% of 50? Choose a reliable method before reaching for a calculator.
Show worked solutionHide worked solution for example 5.01
Recognize. Half of 18 is 9, because taking 18% of 50 is the same product as taking 50% of 18.
Work. 0.18(50) = 18/2 = 9. Another mental route is 20% of 50 minus 2% of 50: 10 − 1 .18*50 is valid, but not necessary here.
Answer: 9.
Check. 10% of 50 is 5 and 20% is 10, so the answer belongs between 5 and 10.
Avoid the trap. Do not confuse 18% with a factor of 18. The percent sign means division by 100, and a sensible bound catches that scale error.
5.02. Solve for the expression, not the variable
If 6x + 9 = 33, what is 2x + 3?
Show worked solutionHide worked solution for example 5.02
Recognize. The requested expression is exactly one-third of the entire left side.
Work. Divide both sides of the given equation by 3: (6x + 9)/3
Answer: 11.
Check. Solving the longer way gives x
Avoid the trap. Divide the entire equation, including the constant 9. Dividing only 6x changes the relationship.
5.03. Add a system when the target is a sum
The equations 3x + 2y = 17 and 2x + 3y = 18 hold. Find x + y.
Show worked solutionHide worked solution for example 5.03
Recognize. Adding the equations produces equal coefficients on x and y.
Work. Adding gives 5x + 5y
Answer: 7.
Check. The individual solution is (x, y)
Avoid the trap. Graphing both lines would work, but it creates two coordinates and an extra addition when the equations already reveal the target directly.
5.04. An equivalent expression needs its domain attached
Simplify (x2 − 16)/(x − 4) and state any restriction from the original expression.
Show worked solutionHide worked solution for example 5.04
Recognize. The numerator is a difference of squares, so algebra exposes the common factor.
Work. Factor x2 − 16
Answer: x + 4, with x ≠ 4.
Check. At x
Avoid the trap. An overlapping graph or matching table values do not restore an excluded input. A complete simplification preserves the original domain.
5.05. Read a vertex instead of hunting for it
For real x, what is the minimum value of f(x) = 4(x − 3)2 − 7?
Show worked solutionHide worked solution for example 5.05
Recognize. A nonnegative square multiplied by a positive number cannot make the function smaller than its constant offset.
Work. Since (x − 3)2 ≥ 0, 4(x − 3)2 − 7 ≥ −7. Equality occurs at x
Answer: −7.
Check. f(3)
Avoid the trap. The input of the minimum is 3; its value is −7. Do not report the vertex's x-coordinate when the question asks for the output.
5.06. Check the case that removes the quadratic term
For which real values of a does ax2 + 6x + 9 = 0 have exactly one distinct real solution?
Show worked solutionHide worked solution for example 5.06
Recognize. There is a quadratic case when a ≠ 0 and a separate linear case when a
Work. For a ≠ 0, a single distinct real root requires discriminant 62 − 4a(9) = 36 − 36a = 0, so a
Answer: a
Check. At a
Avoid the trap. A discriminant-only solution would miss a
5.07. Use an intersection to tame awkward coefficients
The system 1.7x + 0.8y = 11.7 and 0.6x − 1.3y = −2.2 has solution (x, y). Find x − y.
Show worked solutionHide worked solution for example 5.07
Recognize. The numerical coefficients make a graph a reasonable candidate-finding method, followed by exact decimal substitution.
Enter: 1.7*x+0.8*y=11.7
Enter: 0.6*x-1.3*y=-2.2
Work. Inspect the intersection in a window covering approximately 0 ≤ x ≤ 8 and 0 ≤ y ≤ 8. The point is (5, 4), so the requested difference is 5 − 4
Answer: 1.
Check. 1.7(5) + 0.8(4) = 8.5 + 3.2 = 11.7 and 0.6(5) − 1.3(4) = 3 − 5.2 = −2.2. The exact candidate satisfies both equations.
Avoid the trap. Do not copy the first coordinate, 5, into the answer field. The graph supplies inputs to the requested calculation; it does not decide which quantity to report.
5.08. Let the calculator do a long power
A model gives P = 1200(1.035)8. Find P to the nearest whole number.
Show worked solutionHide worked solution for example 5.08
Recognize. The setup is already complete, and the remaining repeated multiplication is suitable for numerical calculation.
Enter: 1200*(1.035)^8
Work. The value is approximately 1580.170844. Rounding only the final result gives 1580. Preserve the full factor 1.035 and the full exponent 8.
Answer: 1580.
Check. The factor exceeds 1, so the value must exceed 1200. The computed ratio P/1200 ≈ 1.317 represents about 31.7% total growth, not 8(3.5%) exactly.
Avoid the trap. Do not replace repeated multiplicative growth with 1200(1 + 8 · 0.035). That is a different, linear model.
5.09. Use a complementary-angle identity
If cos 28° = k, what is sin 62° in terms of k?
Show worked solutionHide worked solution for example 5.09
Recognize. The angles are complementary because 28 + 62
Work. For complementary acute angles, sin θ
Answer: k.
Check. In a right triangle, a side opposite one acute angle is adjacent to the other, while the hypotenuse is shared. The two ratios are equal.
Avoid the trap. A rounded decimal approximation loses the exact relationship requested in terms of k. Recognize the identity before choosing a calculator mode.
5.10. A continuous threshold needs an integer finish
A quantity is modeled by P(n) = 200(1.12)n, where n is a nonnegative integer. What is the least n for which P(n) > 400?
Show worked solutionHide worked solution for example 5.10
Recognize. The model increases with n, so adjacent integer values around the threshold can establish the first success.
Enter: 200*(1.12)^6
Enter: 200*(1.12)^7
Work. P(6) ≈ 394.765 < 400 and P(7) ≈ 442.136 > 400. Since the growth factor is greater than 1, earlier integer inputs produce smaller values. Therefore 7 is the first integer meeting the strict inequality.
Answer: 7.
Check. The two neighboring evaluations show both feasibility and minimality. A graph can locate the transition, but it does not remove the integer condition.
Avoid the trap. Do not round a continuous crossing to the nearest integer without testing it. The first value strictly above a threshold may require moving to the next integer.
5.11. Use coefficients when the claim holds for every input
The identity (ax + b)(x − 3) = 2x2 − 5x − 3 holds for all real x. Find a + b.
Show worked solutionHide worked solution for example 5.11
Recognize. An identity requires corresponding polynomial coefficients to match.
Work. Expanding gives ax2 + (b − 3a)x − 3b. Matching the x2 coefficients yields a
Answer: 3.
Check. The middle coefficient becomes b − 3a = 1 − 6 = −5, matching the given polynomial as required.
Avoid the trap. Regression on selected inputs is unnecessary and can hide whether a statement is an identity. Matching every coefficient proves equality for every real input.
5.12. A fitted relationship is not a causal conclusion
In an observational study of volunteers, students who report more study-app use tend to have higher scores. A fitted line has positive slope. Does this establish that the app caused higher scores?
Show worked solutionHide worked solution for example 5.12
Recognize. The question concerns study design, not the arithmetic of the regression.
Work. No. The reported data show an association in the observed group, but app use was not randomly assigned. Motivation, prior preparation, or other differences could contribute to both app use and scores. Volunteer selection also limits any automatic generalization to all students.
Answer: No; the described evidence supports an observed association, not a demonstrated causal effect.
Check. Ask what the study controlled or randomly assigned. A more precise fitted slope does not answer that design question.
Avoid the trap. Neither a steep slope nor a strong fit alone establishes causation. Use the study's design to judge the claim, not the number of digits in the output.
5.13. Rebalance the last five minutes — Workflow scenario
You have five minutes left, three unanswered questions, and two answered questions marked for review. Describe a sensible plan rather than a guaranteed optimal schedule.
Show worked solutionHide worked solution for example 5.13
Recognize. Unanswered and marked are different statuses. Protect opportunities to record answers before spending all remaining time polishing one response.
Work. Use the question menu to identify blanks. One possible budget is 120 seconds for the most approachable unanswered work, 90 seconds for targeted review, 60 seconds for remaining responses and entry checks, and a 30-second buffer. The total is 300 seconds. Adjust immediately if a task takes less time or stalls.
Answer: A flexible plan that prioritizes remaining responses, then specific review, and includes an entry audit.
Check. Confirm that each intended choice is selected or each numeric response is typed. A mark-for-review flag does not fill an answer field.
Avoid the trap. This allocation is an original planning example, not a College Board rule or a proven optimal strategy. Do not keep following a schedule that is clearly failing to fit the remaining work.
5.14. Distinguish a calculator mistake from an app failure — Workflow scenario
During a practice rehearsal, compare two problems: an expression has a syntax error, and the entire testing app stops responding. What should your recovery plan distinguish?
Show worked solutionHide worked solution for example 5.14
Recognize. A mathematical entry problem can often be repaired within the calculator. A testing-device problem requires the official support process.
Work. For the syntax error, inspect grouping, function spelling, and definitions, or use a valid algebraic route. For an app failure during an actual test, raise your hand and follow the proctor's and app's instructions. Do not independently open a browser calculator, use a phone, or assume the timer paused.[20]
Answer: Repair the expression for an input error; alert the proctor for a testing-app or device failure.
Check. The response must preserve testing rules as well as mathematical progress. A malfunction does not authorize outside assistance or an unapproved device.
Avoid the trap. Repeatedly changing a formula cannot fix an unresponsive app, and a wrong graph window is not evidence that the device has failed. Diagnose the category first.
5.15. Combine modeling, graphing, integer checks, and entry
A quantity is modeled by V(t) = 1200(0.8)t/3 for t ≥ 0, measured in days. What is the first whole-number day on which V(t) < 500?
Show worked solutionHide worked solution for example 5.15
Recognize. The quantity is multiplied by 0.8 every three days. The requested input is an integer satisfying a strict inequality.
Enter: y=1200*(0.8)^(x/3)
Enter: y=500
Enter: 1200*(0.8)^(11/3)
Enter: 1200*(0.8)^(12/3)
Work. A graph suggests a crossing between 11 and 12 days. Direct evaluation gives V(11) ≈ 529.47 > 500 and V(12) = 491.52 < 500. The model decreases for t ≥ 0, so no earlier whole-number day works.
Answer: Enter 12.
Check. At day 12, four three-day periods have passed: 1200(0.8)4
Avoid the trap. Do not report 500, which is an output, or the approximate continuous crossing, which is not a whole-number day. A complete solution ends with the requested input in the response field.
Complete the final transfer
Stage | What it establishes |
|---|---|
Model | 0.8 is the multiplier per three days. |
Graph | The continuous crossing lies between days 11 and 12. |
Integer checks | Day 11 fails; day 12 succeeds. |
Monotonicity | Earlier days have larger values, proving minimality. |
Response | The requested whole-number input is |
Transfer habit
After a successful calculator step, ask what remains. Here the remaining work is not more precision: it is interpreting a strict inequality, choosing an integer, and entering the requested input.
Section synthesis: finish the complete chain
Choose a method from the mathematical structure, not from habit. Let the calculator handle appropriate computation, but keep responsibility for units, conditions, and interpretation. A complete solution includes a recorded response and, where useful, a review flag before the module boundary.
5.6 Section exit check
Work without the lesson open. Choose a method, show enough reasoning to check the result, and record any tool or interpretation error. These questions include tool-decision drills as well as mathematical problems.
Before checking the solutions
Name one question where a manual method was shorter, one place where calculator setup needed care, and one check that would catch a plausible wrong answer. If you cannot name an independent check, return to the relevant worked example.