SAT Problem-Solving and Data Analysis: Ratios, Statistics, and Evidence

SAT Problem-Solving and Data Analysis: Cover, diagnose, choose, practice, verify

SAT Problem-Solving and Data Analysis covers ratios, rates, proportional relationships, units, percentages, one- and two-variable data, probability and conditional probability, sampling, margin of error, and evaluating statistical claims. Preserve meaning before arithmetic: name the context, unit, comparison base, population or condition, and the conclusion the design can support. The Digital SAT Math guide owns section structure and the four-domain overview. Linear models belong to Algebra, nonlinear structure to Advanced Math, and spatial measurement, triangles, circles, and trigonometry to Geometry and Trigonometry—even when a story is attached.

Diagnose the first meaning failure: unit or conversion, percent or conditional base, display reading, sample or population confusion, or a claim stronger than the design. Choose the method only after the quantity is defined—unit rate, correct base, distribution feature, residual interpretation, or design-limited inference. Practice with the Student Question Bank filtered to Problem-Solving and Data Analysis and the diagnosed skill. Verify improvement on a later mixed Bluebook checkpoint by counting fewer repeated denominator, unit, and evidence-boundary errors rather than forecasting a score gain.

Cover the official Problem-Solving and Data Analysis domain first

College Board places ratios, rates, units, percentages, data distributions, scatterplots, probability, sampling, margin of error, and statistical claims inside Problem-Solving and Data Analysis. These topics share a requirement: the numbers must stay connected to what was measured, compared, sampled, or inferred.

This page owns quantitative interpretation only. Section structure and the four-domain overview stay on the Digital SAT Math guide. Algebra owns linear relationships, Advanced Math owns nonlinear expressions and functions, and Geometry and Trigonometry owns spatial relationships. A story or graph does not assign the domain by itself.

Use four questions to classify the work: What quantity is being compared? What unit or denominator defines it? What group or interval does the evidence describe? What conclusion does the design support? Those questions prevent correct arithmetic from becoming a wrong answer.

Choose rate and unit methods that keep labels attached

A ratio compares two quantities; a rate compares quantities with different units; a unit rate fixes the denominator at one. Write both labels before calculating. The expression 180 divided by 3 is not meaningful until it becomes 180 kilometers divided by 3 hours, or 60 kilometers per hour.

Original teaching example: a printer uses 14 milliliters of ink for 35 pages. The unit rate is 14 divided by 35, or 0.4 milliliter per page. For 120 pages at the same constant rate, multiply 0.4 by 120 to get 48 milliliters. The proportional assumption and the final unit are part of the answer.

Conversion chains should cancel units visibly. If a value is given per minute and the question asks per hour, multiply by 60 minutes per hour so minutes cancel. Estimation catches inverted conversions: an hourly amount should generally be larger than the matching per-minute amount, not smaller.

Choose the percent base before any formula

Every percentage has a base. Before using a formula, underline the quantity that represents 100 percent. Percent of a value, percent increase, percent decrease, reverse percentage, and percentage-point change use related arithmetic but answer different questions.

Original teaching example: a membership rises from 320 to 368. The increase is 48, and the original base is 320, so the percent increase is 48 divided by 320, or 15 percent. Returning from 368 to 320 is a decrease of 48 divided by 368, about 13 percent. The same absolute change produces different percentages because the bases differ.

Use multipliers to keep direction visible: a 15 percent increase multiplies by 1.15, while a 15 percent decrease multiplies by 0.85. Percentage points compare two rates directly. If a survey result moves from 42 percent to 47 percent, that is a 5-point increase but about a 12 percent relative increase.

Choose distribution evidence, not a single average habit

A distribution is more than its average. Inspect center, spread, shape, and unusual values. Mean uses every value and is sensitive to extremes; median depends on order and is often more stable when one value is far from the rest. Range measures total span, while other measures describe different kinds of spread.

Original teaching example: five delivery times are 18, 19, 20, 21, and 42 minutes. The median is 20, while the mean is 24. The 42-minute value pulls the mean upward but does not change the middle position much. If a sixth ordinary time is added, recompute from the full set rather than adjusting by intuition.

When comparing two groups, do not decide from center alone. One group can have a higher median and much wider spread. A box plot, histogram, or table may reveal overlap, skew, or an outlier that changes the practical conclusion. State exactly which feature supports the comparison.

Choose association methods that respect evidence limits

A scatterplot shows paired values and can suggest direction, form, and strength of association. Positive association means larger values of one variable tend to accompany larger values of the other; negative association means they tend to move in opposite directions. Neither statement proves causation.

Original teaching example: a study log records weekly practice minutes and typing accuracy for 24 learners. The points rise overall but show substantial scatter. A fitted line predicts 86 percent accuracy at 150 minutes. That prediction describes the model, not a guarantee for every learner, and using it far beyond the observed minutes would be extrapolation.

A residual is observed value minus predicted value. Positive residuals lie above the model's prediction and negative residuals lie below it. Look for pattern: random scatter around zero supports the form more than a curved residual pattern does. Always interpret slope and intercept with units and within a meaningful range.

Choose probability and claim methods from the design

Probability begins with a defined sample space. For equally likely outcomes, divide favorable outcomes by total outcomes. Conditional probability changes the denominator to the group satisfying the condition. Write the condition first so the denominator does not silently remain the whole table.

Sampling and assignment answer different questions. Random sampling helps a sample represent a population, so a result may generalize. Random assignment helps create comparable treatment groups, so an experiment can support a causal conclusion. A large convenience sample can still be biased, and an observational association can still have other explanations.

Margin of error describes uncertainty around an estimate under the stated sampling method. Larger random samples generally reduce sampling variability, but they do not repair biased selection or leading questions. When evaluating a claim, identify the population, sample method, assignment method, measured outcome, and exact conclusion before deciding what the evidence supports.

Diagnose the data-analysis error pattern first

Separate a wrong calculation from a wrong interpretation before assigning drills. Review the first decision that failed: choosing the denominator, preserving units, identifying the percent base, reading a display, defining the sample, or making a conclusion stronger than the design permits.

Use College Board domain feedback as a starting signal, then review individual official questions. A broad domain indicator cannot tell you whether the recurring problem is rates, distributions, probability, or inference. Include correct guesses when you cannot explain the denominator or conclusion reliably.

Record the original cue, the selected quantity or evidence, the corrected interpretation, and a next task. After several sets, count recurring codes. A repeated denominator error needs conditional-probability and percent-base work; a repeated evidence error needs design and claim practice, not faster arithmetic.

• •UNI - a unit, conversion, or rate direction changed.
• •BAS - the denominator or percent base was wrong.
• •DSP - a table, graph, or distribution feature was misread.
• •SAM - the sample, population, or condition was confused.
• •CLM - the conclusion exceeded the design or observed evidence.

Practice with official data-analysis tools and verify transfer

Start from reviewed Bluebook evidence or an official set. Name the Problem-Solving and Data Analysis skill and error code before adding volume. In the Student Question Bank, filter Math to the domain, choose the relevant skill and difficulty, and explain units, base, population, and conclusion before calculating.

Use Khan Academy when the underlying ratio, statistical, or probability method is missing. Then return to unseen official questions and mix the domain's task families. Interleave the three sibling Math domains so a story or graph cannot force the wrong owner.

Verify improvement after the reasoning changes. Compare denominator choices, unit checks, and evidence boundaries across similar conditions rather than forecasting a point gain. If you want private SAT exam assistance after you have a real data-analysis error log, use the booking page or the take-my-SAT service page. Pay-after-pass is billing only, not a score guarantee.