Checking units, constraints, and reasonableness

5.1 Audit the model, the computation, and the response

A solution can fail at three different levels. The model may describe the wrong relationship; the computation may mishandle a correct model; or the final response may report the wrong quantity. Effective checking distinguishes these failures instead of simply repeating the same arithmetic.

Layer

Main question

Typical failure

Model

Did I translate the original conditions correctly?

Using the wrong percent base or counting too few fence segments.

Computation

Do the steps follow correctly from the model?

Dropping a negative sign or rounding too early.

Response

Does the answer report the requested quantity and format?

Reporting x instead of x2, radius instead of diameter, or a decimal rate instead of a percent.

Make the check independent

Substitute a proposed solution into the original equation. Rebuild a total from its parts. Use a different representation when it is quick: compare an algebraic root with the graph’s intercept, or a numerical answer with a rough bound. Repeating the same steps can repeat the same mistake.

A plausibility check is a filter, not usually a proof. A probability of 1.4 is impossible. A probability of 0.4 may be possible but still incorrect. A negative physical width is invalid; a positive width still needs to satisfy the original dimensions and area.

The final five questions

What quantity did I find? In what units? Which original conditions does it satisfy? Is its sign and size plausible? Did I enter that quantity rather than an intermediate result?

Match the depth of the check to the risk

A one-step arithmetic result may need only a sign-and-size check. A radical equation needs an original-equation check. A parameter question needs boundary cases. A two-way-table probability needs a denominator check. Choose the check that targets the likely failure rather than redoing everything indiscriminately.

A check can justify changing an answer

An identified contradiction with the original equation is evidence. A vague feeling that an answer “looks too simple” is not. Change a response when a verified correction improves it, not merely because a second method produces an unexplained different number.

5.2 Use units as algebraic information

Units can reveal which operation is needed. If the target is a distance and the given quantities are speed and time, multiplication produces the right units. If the target is a rate, divide an amount by the corresponding interval.

A conversion factor is a ratio equal to 1. Arrange it so that unwanted units cancel:

kilometershour · 1000 meters1 kilometer · 1 hour3600 seconds = meterssecond.

Square or cube the conversion when the quantity requires it

Since 1 m = 100 cm, 1 m2 = 10,000 cm2, 1 m3 = 1,000,000 cm3. The same principle applies to similarity: a length multiplier s becomes an area multiplier s2 and a volume multiplier s3. Do not apply a linear factor to a squared or cubed quantity.

Check

What it can catch

Units of a slope

Dollars per item versus items per dollar; meters per second versus meters.

Units of an exponent

The exponent in a growth model must be a pure number; time units must match the growth interval.

Area versus length

A reported side length may need to be squared or multiplied by another length.

Radius versus diameter

A factor-of-two error becomes a factor-of-four error in circle area.

Percent versus proportion

0.25 and 25% represent the same rate but different numerical response scales.

Do not add incompatible quantities

You may add 3 hours and 45 minutes after converting them to a common unit. You may not add 3 hours and 45 miles to produce a meaningful total. In a cost model, every additive term must have units of dollars; a rate in dollars per item must first be multiplied by an item count.

A limitation of unit checks

Two wrong formulas can have the same units. Units may show that an operation is impossible without uniquely determining the correct model. Use them alongside the relationships in the problem, not instead of those relationships.

5.3 Preserve domains, integer limits, and feasible regions

A mathematical domain states which inputs make the expression meaningful. A contextual domain states which inputs make sense in the situation. The final solution must belong to both.

Mathematical restrictions: denominators are nonzero; square-root radicands are nonnegative; an equation involving a square root retains the root’s sign; a logarithm, when used as an optional calculation, requires a positive argument.

Contextual restrictions: lengths and durations may need to be positive; counts are nonnegative integers; a probability lies between 0 and 1; a model may be intended only for a specified interval.

Transformations can hide restrictions

Multiplying by a denominator does not make its zero an allowed input. Squaring an equation can introduce roots. Canceling a factor can remove a visible zero from a denominator while leaving the original exclusion in force. Keep these restrictions beside the work and revisit them at the end.

Feasible integer decisions are not ordinary rounding

For enough buses to transport everyone, round capacity upward. For the largest whole number of items within a budget, round the allowable count downward. For a strict inequality, an integer exactly at the boundary may be excluded.

The best check is neighboring-integer verification: show that the chosen integer works and that the next relevant integer fails. This establishes minimum or maximum feasibility more clearly than a rounded decimal alone.

A number line has an open boundary at 7.5 and marks the region to its left. The adjacent integers 7 and 8 are labelled. For integer n less than 7.5, the greatest allowed n is 7.78boundary7.5Forintegern<7.5,thegreatestallowednis7.

Constraints combine by intersection

An answer must satisfy every applicable condition. A dimension that fits the budget but exceeds the area limit is invalid. A root that satisfies the squared equation but has an impossible original sign is invalid.

5.4 Use bounds and distinguish certainty from estimation

A bound is information even when it does not determine an exact value. Positive quantities with a fixed sum can have different products, but the product may still have a maximum. An average of positive-duration speeds must lie between the speeds, but the correct average still depends on the durations.

Build a range before chasing an exact answer

For a positive rectangle with perimeter 40, let its width be w and its length be 20 − w. Then

A = w(20 − w) = 100 − (w − 10)2 ≤ 100.

The area need not be 100; that value occurs only for a square. Distinguish “must be at most” from “must equal.”

Retain precision until the final decision

An approximation is useful only at the resolution needed by the question. Keep exact fractions and radicals where practical. Round once at the end when the question requests a rounded value. For a threshold or integer maximum, test nearby allowed values rather than trusting a rounded boundary.

A calculator display very near zero can result from rounding. A graph window that shows one crossing does not prove no other crossing exists. Use exact algebra or original-equation verification when the conclusion depends on equality or solution count.

Check the strength of a statistical statement

A sample proportion can be used to estimate a population count under an appropriate sampling design. It does not establish that exact count with certainty. A numerical association alone does not establish a cause. These distinctions belong to the SAT’s sample-inference and statistical-claims skill areas.[9]

For pooled data, reconstruct counts or weighted totals before combining rates. Two group percentages should not be averaged equally unless the groups deserve equal weight for the target being computed. A large group’s rate contributes more to an overall individual-level rate than a small group’s rate does.

Section 5 exit standard

You can identify what a check proves and what it does not. You reject impossible answers, preserve original restrictions, handle discrete thresholds, and distinguish an exact result from an estimate or a bound.

15 worked examples

5.01. Cancel units through a rate conversion

A vehicle travels at 72 kilometers per hour. What is this speed in meters per second? Use 1 km = 1000 m and 1 hour = 3600 seconds.

Show worked solutionHide worked solution for example 5.01

Recognize. Convert both the distance and time units. Arrange the factors so the original units cancel.

Work. 72kmh · 1000 m1 km · 1 h3600 s = 20ms.

Answer: 20 meters per second.

Check. In 3,600 seconds, a speed of 20 meters per second covers 72,000 meters, or 72 kilometers.

Avoid the trap. Multiplying by 3,600 instead of dividing reverses the time conversion. Write the units in the factors to make their cancellation visible.

5.02. Apply a conversion factor to area, not just length

A panel has area 0.75 square meter. What is its area in square centimeters? Use 1 m = 100 cm.

Show worked solutionHide worked solution for example 5.02

Recognize. Both dimensions of an area are affected by the length conversion.

Work. Since 1 m2 = (100 cm)2 = 10,000 cm2, the panel’s area is 0.75(10,000) = 7,500 square centimeters.

Answer: 7,500 square centimeters.

Check. A 75-centimeter by 100-centimeter rectangle has area 7,500 square centimeters and also 0.75 m × 1 m = 0.75 m2.

Avoid the trap. Multiplying by 100 gives 75, which is a linear conversion applied to a squared quantity. The exponent on the unit determines the exponent on the conversion factor.

5.03. Use the correct base when reversing a change

A price rises from $80 to $100. By what percent must the new price decrease to return to $80?

Show worked solutionHide worked solution for example 5.03

Recognize. The requested decrease is measured relative to the new price, $100.

Work. The required decrease is $20, so the percent decrease is 100(20/100) = 20%.

Answer: 20%.

Check. Decreasing $100 by 20% gives $80. The original increase was 20/80 = 25%, which has a different base.

Avoid the trap. Equal dollar changes in opposite directions do not generally have equal percentage changes. State the denominator in words before dividing.

5.04. Check which group belongs in the denominator

The table shows 100 students. A student who walks to school is selected at random. What is the probability that this student is in grade 11?

Grade

Walks

Does not walk

Total

Grade 10

18

42

60

Grade 11

12

28

40

Total

30

70

100

Show worked solutionHide worked solution for example 5.04

Recognize. Selection is restricted to the 30 students who walk.

Work. Of these 30 students, 12 are in grade 11. The probability is 12/30 = 2/5.

Answer: 2/5.

Check. The numerator is a subset of the denominator, and the result lies between 0 and 1.

Avoid the trap. 12/40 reverses the condition: it is the probability of walking among grade 11 students. 12/100 ignores the restriction to walkers.

5.05. Filter a geometric root and then report the requested side

A rectangle’s length is 5 centimeters greater than its width, and its area is 84 square centimeters. What is its length?

Show worked solutionHide worked solution for example 5.05

Recognize. Let w be the width. The model is w(w + 5) = 84, and both side lengths must be positive.

Work. Factoring gives w2 + 5w − 84 = (w + 12)(w − 7) = 0. Reject w = −12. The width is 7, so the requested length is 7 + 5 = 12.

Answer: 12 centimeters.

Check. The dimensions are 7 and 12, their difference is 5, and their product is 84.

Avoid the trap. Reporting 7 stops at the variable you chose rather than the quantity requested. Reporting a negative root fails the geometric domain even though it satisfies the polynomial equation.

5.06. Verify every candidate after squaring

Solve x + 6 = x over the real numbers.

Show worked solutionHide worked solution for example 5.06

Recognize. The right side must be nonnegative, so x ≥ 0. Squaring can introduce candidates that violate this sign restriction.

Work. Squaring gives x + 6 = x2, so (x − 3)(x + 2) = 0. The candidates are 3 and −2. In the original equation, x = 3 gives 3 = 3, while x = −2 gives 2 = −2.

Answer: x = 3.

Check. The accepted solution satisfies both the original radicand condition and the nonnegative right-side condition.

Avoid the trap. The radicand at x = −2 is positive, but that alone does not make it a solution. The two original sides must also have equal signs and values.

5.07. Recognize when the only candidate is forbidden

How many real solutions does x + 1x − 2 = 3x − 2 have?

Show worked solutionHide worked solution for example 5.07

Recognize. The original equation excludes x = 2. Multiplying by the denominator is valid only for allowed inputs.

Work. For x ≠ 2, multiplication by x − 2 gives x + 1 = 3, so the only candidate is x = 2. It is excluded. Therefore there are no real solutions.

Answer: 0.

Check. At every allowed input, subtracting the right side from the left gives (x − 2)/(x − 2) = 1, never 0.

Avoid the trap. A candidate produced after clearing denominators is not automatically valid. The original restriction survives every equivalent transformation on its domain.

5.08. Convert a growth multiplier into a percent increase

A quantity grows by 20% at the end of each year. What is its total percent increase after three years?

Show worked solutionHide worked solution for example 5.08

Recognize. The multiplier after three years is 1.23. The percent increase is based on the amount above the original, not the final multiple itself.

Work. The final multiplier is 1.23 = 1.728. Subtract the original multiplier 1 and convert to a percent:

100(1.728 − 1) = 72.8%.

Answer: 72.8%.

Check. Starting from 100 gives 120, 144, and 172.8 after the successive years, an increase of 72.8.

Avoid the trap. The final amount is 172.8% of the original, not a 172.8% increase. Adding 20 + 20 + 20 also ignores the changing base.

5.09. Check the angle unit and a geometric bound

In a right triangle, an acute angle measures 35°. The leg adjacent to this angle is 12 centimeters. What is the opposite leg’s length, to the nearest tenth of a centimeter?

Show worked solutionHide worked solution for example 5.09

Recognize. Tangent relates opposite to adjacent. The stated angle is in degrees, so numerical evaluation must use degrees.

Work. If the opposite leg is h, then tan 35° = h/12, giving h = 12 tan 35° ≈ 8.4025.

Answer: 8.4 centimeters.

Check. Because 35° < 45°, the opposite leg is shorter than the adjacent leg. Thus 0 < h < 12, consistent with the answer.

Avoid the trap. A calculator set to radians evaluates a different angle. A correct formula entered with the wrong angle unit is still the wrong computation. The bound helps flag a suspicious result but does not replace checking the mode.

5.10. Use a neighboring-integer check for capacity

A school must transport 275 students. Each bus can carry at most 48 students. What is the least number of buses needed?

Show worked solutionHide worked solution for example 5.10

Recognize. Capacity must cover every student. The count of buses is a positive integer satisfying 48b ≥ 275.

Work. The quotient 275/48 ≈ 5.729 shows that five buses are insufficient. Round upward to six buses.

Answer: 6 buses.

Check. Five buses carry at most 240 students, leaving 35 without seats. Six buses carry up to 288, which is enough.

Avoid the trap. Ordinary rounding is not the underlying rule. The required direction comes from the capacity inequality; the same kind of decimal in a maximum-budget problem may need to be rounded downward.

5.11. Separate a population estimate from an exact count

A simple random sample of 240 students from a school of 1,000 contains 150 students who prefer a proposed schedule. Based on the sample proportion, what is the estimated number of students in the school who prefer the schedule?

Show worked solutionHide worked solution for example 5.11

Recognize. The task asks for an estimate using a sample proportion, not a census count.

Work. The sample proportion is 150/240 = 5/8 = 0.625. Applying it to the school population gives 0.625(1,000) = 625.

Answer: An estimated 625 students.

Check. The sample represents 62.5% support, so the corresponding estimate should exceed half of 1,000 and be less than 1,000.

Avoid the trap. The calculation does not prove that exactly 625 students prefer the schedule. It also does not justify applying the estimate to a different school’s students. The numerical estimate and the strength of the conclusion must be kept distinct.

5.12. Pool counts rather than averaging group rates equally

In one group, 10 of 20 students completed a task. In another group, 45 of 180 students completed it. What percent of the combined 200 students completed the task?

Show worked solutionHide worked solution for example 5.12

Recognize. The target is the completion rate among individuals in both groups. The group sizes are unequal, so the two percentages need different weights.

Work. The total number of completions is 10 + 45 = 55, out of 20 + 180 = 200 students. Thus the combined rate is 55/200 = 0.275 = 27.5%.

Answer: 27.5%.

Check. The group rates are 50% and 25%. The combined rate is much closer to 25% because the second group contains nine times as many students.

Avoid the trap. The unweighted average (50% + 25%)/2 = 37.5% gives equal weight to groups rather than to students, so it answers a different question.

5.13. Verify the trigonometric ratio, not just the missing side

Triangle ABC is right at C. The hypotenuse AB is 13 and the leg AC is 5. What is tan A?

Show worked solutionHide worked solution for example 5.13

Recognize. The unknown leg is opposite A, and AC is adjacent to A. Tangent uses the two legs, not the hypotenuse.

Work. The Pythagorean theorem gives BC = 132 − 52 = 144 = 12. Therefore tan A = BC/AC = 12/5.

Answer: 12/5.

Check. The opposite leg is longer than the adjacent leg, so A > 45° and its tangent exceeds 1. The result 12/5 is consistent.

Avoid the trap. 12/13 is sin A, not tan A. Correctly finding the missing side does not guarantee that the final ratio answers the question.

5.14. Distinguish a bound from a fixed value

A rectangle has positive side lengths and perimeter 40. Which statement about its area A must be true? A) A = 100 B) A ≥ 100 C) A ≤ 100 D) A = 96

Show worked solutionHide worked solution for example 5.14

Recognize. A fixed perimeter does not fix the area. Express the area with one variable and find a bound.

Work. If one side is w, the other is 20 − w, where 0 < w < 20. Then A = w(20 − w) = 100 − (w − 10)2 ≤ 100.

Answer: C, A ≤ 100.

Check. A 10 by 10 square gives area 100, while an 8 by 12 rectangle gives 96. Both satisfy the perimeter, disproving a single fixed area.

Avoid the trap. Choosing one convenient rectangle can show what is possible, not what must happen for every rectangle. The squared expression establishes the universal upper bound.

5.15. Combine a discrete dimension with two feasibility limits

A rectangular garden has positive integer width w meters and length w + 3 meters. Its area may not exceed 150 square meters. Fencing all four sides costs $7 per meter, with a budget of at most $350. What is the greatest possible garden area?

Show worked solutionHide worked solution for example 5.15

Recognize. The width must satisfy both a budget inequality and an area inequality. The target is the area, not the width. Start with the simpler constraint.

Work. The perimeter is 2w + 2(w + 3) = 4w + 6, so the budget gives

7(4w + 6) ≤ 350 w ≤ 11.

The area condition is w(w + 3) ≤ 150. Test the largest width allowed by the budget: at w = 11, the area is 11(14) = 154, which is too large. At w = 10, the area is 10(13) = 130, which is allowed. For positive w, both w and w + 3 increase as w increases, so the area increases. Thus the greatest feasible area occurs at w = 10.

Answer: 130 square meters.

Check. The dimensions 10 and 13 have perimeter 46, so fencing costs 46(7) = 322 dollars. The area 130 satisfies its limit. The next width, 11, fits the fencing budget exactly but violates the area limit.

Alternative. Solving w2 + 3w − 150 ≤ 0 gives the positive boundary (−3 + 609)/2 ≈ 10.839. The greatest feasible integer is 10, not 11. Testing adjacent integers is enough once the simpler budget restricts the candidates.

Avoid the trap. Checking only one constraint would accept an invalid garden. Rounding a real boundary to the nearest integer would also fail. Finally, entering 10 would report the width instead of the requested area.

A complete final check

The accepted result has the right mathematical domain, satisfies every contextual limit, is optimal among permitted choices, and reports the requested quantity with its units.

5.6 Section exit check

For each result, write the most relevant independent check.

Topic practice

Review prompt

For E5.2, check the original sign. For E5.3 and E5.5, check neighboring integers. For E5.4, check the denominator and the weighting.