SAT · Math · Unit 03 Problem-Solving and Data AnalysisProblem-Solving and Data Analysis: Unit Study Guide
03.01 Ratios, Rates, Proportions and Units
Problem-Solving and Data Analysis is about 15% of SAT Math (5–7 of 44 questions). This topic covers ratios, rates and proportions: unit rates, scale drawings, sharing in a ratio, direct and inverse variation, unit conversions (including square and cubic units), and rate problems such as distance, work and mixtures.
Ratios, Rates and Proportions
A ratio compares quantities; a proportion sets two ratios equal and is solved by cross-multiplying. Unit rates divide by the second quantity.
Key formulas| Proportion | ba=dc⇒ad=bc |
| Unit rate | 1 unitamount |
| Scale | actual=map×scale factor |
Proportional relationships — constant from a graph
Method. Read any labeled point (x, y) on the line and divide y by x.
Worked exampleThe graph shows a proportional relationship y = kx. What is the value of k?

AnswerB — 2
Step-by-step- The line passes through the origin and the point (2, 4).
- k = y ÷ x = 4 ÷ 2 = 2.
Quick check2 × 2 = 4, so the point (2, 4) fits y = 2x.
Common mistakes- A (21): divided x by y instead of y by x
- C (4): gave the y-value of the point
- D (8): multiplied the coordinates
Unit conversions with square and cubic units
Method. Square units use the factor squared (1002=10,000); cubic units use it cubed (1003=1,000,000).
Worked exampleHow many square centimeters are in 5 square meters?
AnswerD — 50,000
Step-by-step- One square meter is 100 × 100 = 10,000 square centimeters, because every dimension is converted.
- So 5 square meters = 5 × 10,000 = 50,000 square centimeters.
Quick check50,000 ÷ 10,000 = 5 square meters, the amount given.
Common mistakes- A (0.0005): divided instead of multiplying
- B (500): multiplied by the length factor 100 only
- C (1,000): multiplied by 100 × 2 instead of 100 to the power 2
Inverse variation — work problems
Method. Multiply the given pair, then divide by the new value.
Worked exampleIt takes 2 workers 20 days to paint a building. Working at the same rate, how many days would it take 10 workers?
AnswerA — 4
Step-by-step- Inverse variation: x × y stays constant.
- 2 × 20 = 40.
- So y = 40 ÷ 10 = 4.
Quick check10 × 4 = 40 = 2 × 20.
Common mistakes- B (12): kept the sum of x and y constant
- C (40): gave the product xy (the constant) instead of y
- D (100): used direct variation instead of inverse variation
Inverse variation
Method. Multiply the given pair, then divide by the new value.
Worked exampley varies inversely with x, and y = 5 when x = 4. What is y when x = 5?
AnswerA — 4
Step-by-step- Inverse variation: x × y stays constant.
- 4 × 5 = 20.
- So y = 20 ÷ 5 = 4.
Quick check5 × 4 = 20 = 4 × 5.
Common mistakes- B (6): added the change in x to y
- C (6.25): used direct variation instead of inverse variation
- D (20): gave the product xy (the constant) instead of y
Dimensional analysis — rate conversions
Method. Multiply by conversion fractions so unwanted units cancel; check every unit is converted.
Worked exampleA tap drips 4 milliliters every minute. How many liters does it drip in a day? (1 liter = 1,000 milliliters)
AnswerC — 5.76
Step-by-step- 4 mL/min × 60 min/h × 24 h/day = 5,760 mL per day.
- ÷ 1,000 = 5.76 liters.
Quick check5.76 L = 5,760 mL, and 5,760 ÷ 1,440 minutes = 4 mL per minute.
Common mistakes- A (0.096): multiplied by 24 but forgot the 60 minutes in an hour
- B (0.24): found liters per hour, not per day
- D (5,760): did not convert milliliters to liters
Sharing in a ratio
Method. Add the ratio numbers, find one part, then multiply.
Worked example$630 is shared among three friends in the ratio 2:5:7. How much is the smallest share?
AnswerB — $90
Step-by-step- Total parts: 2 + 5 + 7 = 14.
- One part = $630 ÷ 14 = $45.
- The smallest share is 2 × $45 = $90.
Quick checkshares $90 + $225 + $315 = $630.
Common mistakes- A ($45): gave the value of one part only
- C ($225): gave the middle share
- D ($315): gave the largest share
Ratios — find the total
Method. Total = boys × (total parts ÷ boy parts) = 16 × 45 = 20.
Worked exampleThe ratio of boys to girls in a club is 4:1. There are 16 boys. How many students are in the club?
AnswerC — 20
Step-by-step- 4 parts are 16 boys.
- So one part is 16 ÷ 4 = 4.
- The club has 4 + 1 = 5 parts: 5 × 4 = 20 students.
Quick check16:4 simplifies to 4:1, and 16 + 4 = 20.
Common mistakes- A (4): found the number of girls only
- B (17): added the girls' ratio number to the number of boys
- D (21): added the ratio total to the number of boys
Multi-step rate problems
Method. Find the per-minute rate first, convert all time to minutes, then multiply.
Worked exampleA printer prints 36 pages every 3 minutes. At this rate, how many pages does it print in 3 hours 20 minutes?
AnswerB — 2,400
Step-by-step- Rate: 36 ÷ 3 = 12 pages per minute.
- Time: 3 × 60 + 20 = 200 minutes.
- Pages: 12 × 200 = 2,400.
Quick check2,400 ÷ 200 = 12 pages per minute = 36 pages per 3 minutes.
Common mistakes- A (2,160): ignored the extra minutes
- C (3,840): treated 3 hours as 300 minutes
- D (7,200): used 36 pages per minute instead of per 3 minutes
Average speed for a round trip
Method. For equal distances, average speed = 2ab/(a + b) = 2 × 20 × 60 ÷ 80 = 30.
Worked exampleA driver travels 180 miles to a lake at 20 miles per hour and drives back the same way at 60 miles per hour. What is the average speed for the whole trip, in miles per hour?
AnswerB — 30
Step-by-step- Time out = 180 ÷ 20 = 9 h.
- Time back = 180 ÷ 60 = 3 h.
- Total distance 360 miles ÷ total time 12 h = 30 mph.
Quick check30 is closer to 20 than to 60, because more time is spent at the slower speed.
Common mistakes- A (15): used the one-way distance with the total time
- C (40): averaged the two speeds
- D (120): used only the time for the return trip
Rate Problems
Distance, work and mixture problems are rate problems: rate × time = amount.
Key formulas| Distance | d=rt |
| Work together | t11+t21=T1 |
| Mixture | c1v1+c2v2=c(v1+v2) |
Distance, rate and time
Method. Distance = rate × time; opposite directions or toward each other → add speeds, same direction → subtract.
Worked exampleTwo cars leave the same town at the same time and drive in opposite directions, one at 32 miles per hour and the other at 37 miles per hour. After how many hours will they be 138 miles apart?
AnswerB — 2
Step-by-step- Moving apart, the distance grows at 32 + 37 = 69 miles per hour.
- Time = distance ÷ rate: 138 ÷ 69 = 2 hours.
Quick check32 × 2 + 37 × 2 = 64 + 74 = 138 ✓.
Common mistakes- A (1): doubled the sum of the speeds
- C (4): divided by the average speed instead of the sum of the speeds
- D (27.6): subtracted the speeds instead of adding
Work problems
Method. Together time = (product)/(sum) of the two times.
Worked exampleLeo can paint a fence in 6 hours. Dev can paint the same fence in 30 hours. Working together at these rates, how many hours will it take them to paint the fence?
AnswerB — 5
Step-by-step- Leo does 61 of the job per hour and Dev does 301.
- Together: 61 + 301 = 51 of the job per hour.
- Time = 1 ÷ (51) = 5 hours.
Quick checkin 5 hours Leo paints 65 and Dev paints 61 of the fence; 65 + 61 = 1 ✓.
Common mistakes- A (0.2): gave the combined rate (jobs per hour) instead of the time
- C (18): averaged the two times
- D (36): added the two times
Mixture problems
Method. Balance the differences: 30 × (60 − 20) = y × (75 − 60).
Worked exampleHow many liters of a 75% acid solution must be added to 30 liters of a 20% acid solution to make a 60% acid solution?
AnswerC — 80
Step-by-step- Let y = liters of the 75% solution.
- Acid: 0.2(30) + 0.75y = 0.6(30 + y).
- Multiply by 100: 600 + 75y = 1800 + 60y.
- Collect: 15y = 1200.
- Divide: y = 80 liters.
Quick checkacid = 6 + 60 = 66 liters, and 60% of 110 = 66 ✓.
Common mistakes- A (16): divided by 75 instead of 75 − 60
- B (30): assumed equal amounts of the two solutions
- D (110): gave the total amount of the mixture
03.02 Percentages
Problem-Solving and Data Analysis is about 15% of SAT Math (5–7 of 44 questions). This topic covers percents: percent of a quantity, percent increase and decrease, finding the original amount, successive changes with multipliers, percent of a percent, and percentage points versus percent change.
Percents
Percent means per hundred. Convert to a decimal, then multiply; percent change compares the change with the original.
Key formulas| Percent of | p% of n=100pn |
| Percent change | oldnew−old×100% |
| Multiplier | increase by r: ×(1+r); decrease by r: ×(1−r) |
| Successive changes | n×multiplier1×multiplier2 |
| Original amount | original=multipliernew |
| Percentage points | 20%→25%: 5 points, a 25% increase |
Percent increase and decrease
Method. New ÷ original = 0.625, so the change is 0.625 − 1 = −0.375 → 37.5%.
Worked exampleThe price of a jacket changed from $296.00 to $185.00. What is the percent decrease?
AnswerA — 37.5%
Step-by-step- Change = $185.00 − $296.00 = $111.00 (a decrease).
- Percent change = change ÷ original × 100 = 111 ÷ 296 × 100 = 37.5%.
Quick check$296.00 − 37.5% of $296.00 = $185.00.
Common mistakes- B (60%): divided by the new value instead of the original
- C (62.5%): gave the new value as a percent of the original
- D (111%): gave the amount of change as the percent
Successive percent changes
Method. Multiply the multipliers: 1.3 × 0.6 = 0.78, then × $100.00 = $78.00.
Worked exampleA price of $100.00 is increased by 30%, and then the new price is decreased by 40%. What is the final price?
A$52.00
B$78.00
C$90.00
D$130.00
AnswerB — $78.00
Step-by-step- Increase: $100.00 × 1.3 = $130.00.
- Decrease that by 40%: $130.00 × 0.6 = $78.00.
Quick checkthe second change is taken from $130.00, not from $100.00, so the result is not $90.00.
Common mistakes- A ($52.00): gave the amount of the second change instead of the final price
- C ($90.00): added and subtracted the percents as if they apply to the same amount
- D ($130.00): applied only the increase
Mixtures — percent concentration
Method. Weighted average: (amount₁ × percent₁ + amount₂ × percent₂) ÷ total amount.
Worked example90 liters of a 15% salt solution are mixed with 60 liters of a 30% salt solution. What is the percent of salt in the mixture?
AnswerA — 21%
Step-by-step- Salt: 0.15 × 90 + 0.3 × 60 = 13.5 + 18 = 31.5 liters, in 150 liters.
- Percent = 31.5 ÷ 150 × 100 = 21%.
Quick check21% lies between 15% and 30%, closer to the percent of the larger amount.
Common mistakes- B (22.5%): averaged the two percents without weighting
- C (24%): matched each percent with the wrong amount
- D (45%): added the two percents
Repeated percent growth
Method. 2,900 × 1.22 = 2,900 × 1.44 = 4,176.
Worked exampleA town has 2,900 people. The population grows by 20% each year. What will the population be after 2 years?
AnswerD — 4,176
Step-by-step- Each year multiply by 1.2: 2,900 → 3,480 → 4,176.
- After 2 years: 4,176.
Quick checkthe growth in year 2 (696) is larger than in year 1 (580), as expected for percent growth.
Common mistakes- A (1,276): gave only the increase
- B (3,480): applied the growth for one year only
- C (4,060): added the same amount each year (simple growth)
Reverse percent — find the original
Method. Divide by the multiplier (0.85); do not take 15% of the new amount.
Worked exampleAfter a 15% discount, a coat costs $190.40. What was the original price?
A$161.84
B$205.40
C$218.96
D$224.00
AnswerD — $224.00
Step-by-step- The new amount is 85% of the original.
- So original = $190.40 ÷ 0.85 = $224.00.
Quick check$224.00 − 15% of $224.00 = $190.40.
Common mistakes- A ($161.84): decreased the new price again
- B ($205.40): added 15 dollars
- C ($218.96): added 15% of the new price
Percentage points versus percent change
Method. Percentage points = simple difference of the rates; percent change = that difference ÷ the original rate × 100.
Worked exampleA bank's interest rate rose from 5% to 6%. By what percent did the rate increase?
AnswerD — 20%
Step-by-step- The rate changed by 1 percentage point.
- Percent change = 1 ÷ 5 × 100 = 20%.
Quick check20% of 5% is 1 percentage point.
Common mistakes- A (0.2%): did not multiply by 100
- B (1%): gave the change in percentage points, not the percent change
- C (11%): added the two rates
03.03 One-Variable Data: Distributions, Center and Spread
Problem-Solving and Data Analysis is about 15% of SAT Math (5–7 of 44 questions). This topic covers one-variable data: mean, median, mode and range, the effect of outliers and of changing values, comparing standard deviation and spread without calculating it, box plots, histograms and dot plots.
Center and Spread
Mean, median and mode describe the center; range, IQR and standard deviation describe the spread. An outlier pulls the mean toward it but barely moves the median. Data spread farther from the mean have a larger standard deviation; the SAT asks you to compare, not to calculate it.
Key formulas| Mean | xˉ=n∑x |
| Sum from the mean | ∑x=nxˉ |
| Median | middle value of the ordered data |
| Range | max−min |
| IQR | Q3−Q1 |
| Add k to every value | center+k, spread unchanged |
Median
Method. Sort first; odd count → middle value, even count → mean of the two middle values.
Worked exampleWhat is the median of the data set 42, 54, 28, 45, 23, 33, 51?
AnswerC — 42
Step-by-step- Order the values.
- So 23, 28, 33, 42, 45, 51, 54.
- There are 7 values.
- The middle one is value 4, which is 42.
- So the median is 42.
Quick check3 values lie on each side of 42 in the ordered list ✓.
Common mistakes- A (4): gave the middle position instead of the value
- B (33): took the value just below the middle
- D (45): did not put the values in order first
Interquartile range
Method. IQR = median of the upper half − median of the lower half.
Worked exampleWhat is the interquartile range (IQR) of 48, 1, 14, 2, 20, 44, 3, 25?
AnswerA — 32
Step-by-step- Order the values.
- So 1, 2, 3, 14, 20, 25, 44, 48.
- The lower half is 1, 2, 3, 14.
- So Q1 = 2.5.
- The upper half is 20, 25, 44, 48.
- So Q3 = 34.5.
- Then IQR = Q3 − Q1 = 34.5 − 2.5 = 32.
Quick check2.5 + 32 = 34.5, the upper quartile ✓.
Common mistakes- B (34.5): gave Q3 only
- C (37): added the quartiles
- D (47): gave the range
Effect of outliers
Method. Find both means, then subtract; an outlier pulls the mean toward it.
Worked exampleThe data set 13, 13, 30, 27, 88, 27 has an outlier. By how much does the mean decrease when the outlier is removed?
AnswerB — 11
Step-by-step- The mean with every value is 198 ÷ 6 = 33.
- The outlier is 88.
- Without it, the mean is 110 ÷ 5 = 22.
- The decrease is 33 − 22 = 11.
Quick check22 + 11 = 33, the original mean ✓.
Common mistakes- A (0): found the change in the median
- C (22): gave the new mean
- D (33): gave the original mean
Comparing data sets
Method. Find the median of each set separately, then subtract.
Worked exampleData set A is 26, 45, 38, 10, 43. Data set B is 15, 34, 26, 27, 27. How much greater is the median of A than the median of B?
AnswerB — 11
Step-by-step- Ordered, A is 10, 26, 38, 43, 45.
- So its median is 38.
- Ordered, B is 15, 26, 27, 27, 34.
- So its median is 27.
- The difference is 38 − 27 = 11.
Quick check27 + 11 = 38, the median of A ✓.
Common mistakes- A (6.6): compared the means instead
- C (16): compared the ranges
- D (33): compared the sums
Weighted mean
Method. Weighted mean = Σ(weight × score) with weights as decimals.
Worked exampleA grade is 50% final exam, 25% projects and 25% classwork. A student scores 69, 71 and 90 in these parts. What is the student's grade?
AnswerB — 74.75
Step-by-step- Multiply each score by its weight.
- So 50% × 69 + 25% × 71 + 25% × 90.
- That is 34.5 + 17.75 + 22.5 = 74.75.
Quick check74.75 lies between the lowest score 69 and the highest score 90 ✓.
Common mistakes- A (69): used only the score with the largest weight
- C (80): matched the weights to the wrong scores
- D (7,475): did not divide by 100
Missing value from a mean
Method. Needed score = target mean × new count − current total.
Worked exampleNina scored 86, 98, 86 on 3 tests. What score on the next test makes the mean of all 4 tests exactly 92?
AnswerC — 98
Step-by-step- For a mean of 92 over 4 tests, the total must be 92 × 4 = 368.
- The first 3 scores add to 270.
- So the next score must be 368 − 270 = 98.
Quick check(270 + 98) ÷ 4 = 92 ✓.
Common mistakes- A (6): used the old number of tests
- B (92): gave the target mean
- D (368): gave the total needed
Standard deviation (concept)
Method. Look for values bunched far from the center on both sides, not just one extreme value.
Worked exampleSet P = {5, 5, 5, 17, 17, 17}, Set Q = {8, 8, 9, 9, 9, 22}, Set R = {21, 22, 23, 25, 26, 27} and Set S = {6, 6, 7, 7, 8, 8}. Which set has the greatest standard deviation?
AnswerC — Set P
Step-by-step- Standard deviation measures how far values typically are from the mean.
- In Set P, the values sit farthest from their mean overall (SD ≈ 6.00).
- The other sets have smaller SDs: Set Q ≈ 5.01, Set R ≈ 2.16, Set S ≈ 0.82.
- So the answer is Set P.
Quick checka large range alone does not make a large SD; Set Q has the greatest range but a smaller SD ✓.
Common mistakes- A (Set Q): chose the set with the greatest range
- B (Set R): chose the set with the greatest mean
- D (Set S): chose the set with the least spread
Distributions and Data Displays
Describe a distribution by its shape, center and spread. A box plot shows the five-number summary; a histogram shows how many values fall in each interval. In a right-skewed distribution the mean is usually greater than the median.
Key formulas| Five-number summary | min, Q1, median, Q3, max |
| IQR | Q3−Q1 |
| Right skew | mean usually greater than median |
| Left skew | mean usually less than median |
| Percentile | percent below x=totalnumber of values below x×100% |
Box plots
Method. Range = end of right whisker − end of left whisker.
Worked exampleThe box plot shows minutes spent reading. What is the range of the data?

AnswerB — 80
Step-by-step- The whiskers end at the minimum, 10, and the maximum, 90.
- The range is 90 − 10 = 80.
Quick check10 + 80 = 90, the end of the right whisker ✓.
Common mistakes- A (70): found the IQR (the length of the box)
- C (90): read the maximum only
- D (100): added the maximum and the minimum
Five-number summary
Method. Sort, split into lower and upper halves, then take the median of each half.
Worked exampleWhat is the first quartile (Q1) of 1, 59, 29, 40, 55, 25, 17, 12, 11, 18, 41, 50?
AnswerB — 14.5
Step-by-step- Order the values.
- So 1, 11, 12, 17, 18, 25, 29, 40, 41, 50, 55, 59.
- The lower half is 1, 11, 12, 17, 18, 25.
- Its median is Q1 = 14.5.
Quick checkQ1 = 14.5 is below the median 27 ✓.
Common mistakes- A (11): took the second value
- C (27): gave the median
- D (45.5): gave Q3 instead
Comparing distributions
Method. Read the same feature from both plots, then subtract.
Worked exampleThe box plots compare two groups. How much greater is the median of Team 1 than the median of Team 2?

AnswerC — 10
Step-by-step- The median of Team 1 is 70 and the median of Team 2 is 60.
- The difference is 70 − 60 = 10.
Quick checkthe answer uses the median line of each plot only ✓.
Common mistakes- A (−10): compared the ranges
- B (5): compared the Q1 values
- D (130): added the medians
Percentiles
Method. Percent below = (number below ÷ total) × 100.
Worked exampleThe data set is 86, 90, 40, 67, 43, 99, 84, 96, 62, 78, 63, 72, 57, 53, 70, 58, 85, 73, 65, 98. What percent of the values are less than 84?
AnswerC — 65%
Step-by-step- Order the values.
- Exactly 13 values are less than 84.
- Out of 20 values, that is 13 ÷ 20 × 100 = 65%.
Quick check13 values below, 1 equal and 6 above make 20 ✓.
Common mistakes- A (13%): gave the count, not the percent
- B (30%): found the percent above it
- D (70%): counted the value itself as below it
Histograms
Method. Find the middle position, then add bar heights from the left until you pass it.
Worked exampleThe histogram shows minutes on homework for 33 students. Each bar includes its left edge but not its right edge. Which interval contains the median?

A40 to 50
B50 to 60
C60 to 70
D70 to 80
AnswerC — 60 to 70
Step-by-step- With 33 values, the median is value number 17.
- The running totals of the bars are 2, 13, 16, 26, 33.
- The total first reaches 17 in the bar from 60 to 70.
Quick checkfewer than 17 values lie below 60 and at least 17 lie below 70 ✓.
Common mistakes- A (40 to 50): chose the tallest bar
- B (50 to 60): chose the middle interval on the scale
- D (70 to 80): miscounted the position of the middle value by a bar
Dot Plots, Bar Graphs and Tables
Count each dot or bar carefully and read the scale before computing a mean or median from a display.
Key formulas| Mean from a frequency display | xˉ=∑frequency∑(value×frequency) |
Bar graphs
Method. Percent change = (new − old) ÷ old × 100; always divide by the earlier year.
Worked exampleThe bar graph shows bikes repaired each year. What was the percent increase from 2020 to 2022?

AnswerC — 150%
Step-by-step- Read the graph.
- So 2020 = 40 and 2022 = 100.
- The change is 60.
- Percent increase = change ÷ original × 100 = 60 ÷ 40 × 100 = 150%.
Quick check40 + 150% of 40 = 40 + 60 = 100 ✓.
Common mistakes- A (40%): gave the original value as a percent of the new value
- B (60%): divided by the new value instead of the original
- D (250%): gave the new value as a percent of the original
03.04 Two-Variable Data: Models and Scatterplots
Problem-Solving and Data Analysis is about 15% of SAT Math (5–7 of 44 questions). This topic covers two-variable data: scatterplots and association, lines and curves of best fit, interpreting the slope and intercept of a model, predicting with a model, residuals, and linear versus exponential models.
Scatterplots and Models
Describe direction and strength of the association. A line of fit ŷ = a + bx predicts y from x: b is the predicted change in y for each 1-unit increase in x, and a is the predicted y when x = 0. Constant percent change calls for an exponential model.
Key formulas| Line of fit | y^=a+bx |
| Residual | y−y^ (actual−predicted) |
| Exponential model | y^=abx |
| Strength | points close to the line: strong; widely scattered: weak |
Scatterplots and association
Method. Compare each point with the line directly above or below it.
Worked exampleThe scatterplot shows drinks sold against temperature (°F) with a line of fit. How many points lie above the line?

AnswerC — 5
Step-by-step- A point is above the line when its y-value is greater than the line's y-value at the same x.
- Checking each of the 8 points, 5 are above and 3 are below.
- So 5 points lie above the line.
Quick check5 above + 3 below = 8 points, and none lie on the line ✓.
Common mistakes- A (3): counted the points below the line
- B (4): counted the points on the right half
- D (8): counted every point
Interpreting a line of fit
Method. Slope = predicted change per 1 unit of x; intercept = prediction at x = 0.
Worked exampleA line of fit for points scored against minutes of practice is ŷ = 4x + 43. Which is the best interpretation of the y-intercept 43?
AFor each additional unit of minutes of practice, the predicted points scored increases by 4.
BWhen minutes of practice is 0, the predicted points scored is 4.
CWhen points scored is 0, the predicted minutes of practice is 43.
DWhen minutes of practice is 0, the predicted points scored is 43.
AnswerD — When minutes of practice is 0, the predicted points scored is 43.
Step-by-step- The y-intercept is the predicted y when x = 0.
- So the intercept means “When minutes of practice is 0, the predicted points scored is 43”.
Quick checkthe chosen sentence uses the right number with x and y in the right roles ✓.
Common mistakes- A (For each additional unit of minutes of practice, the predicted points scored increases by 4.): described the slope instead of the intercept
- B (When minutes of practice is 0, the predicted points scored is 4.): used the slope as the starting value
- C (When points scored is 0, the predicted minutes of practice is 43.): reversed the roles of x and y
Residuals
Method. Residual = actual − predicted; positive means the point is above the line.
Worked exampleA line of fit is ŷ = 0.5x + 13. A data point is (9, 14). What is the residual for this point?
AnswerA — −3.5
Step-by-step- The predicted value is ŷ = 0.5 × 9 + 13 = 17.5.
- Residual = actual − predicted = 14 − 17.5 = −3.5.
Quick checkthe point is below the line, matching the negative residual ✓.
Common mistakes- B (3.5): subtracted in the wrong order (predicted − actual)
- C (14): gave the actual value
- D (17.5): gave the predicted value
Predicting with a Line of Best Fit
Substitute the x-value into the equation of the line of best fit; the result is a prediction, not an exact value.
Key formulasPredicting with a line of best fit
Method. Plug the x-value into the regression equation.
Worked exampleA line of best fit for data relating x (months) and y (sales in thousands of dollars) is ŷ = 5x + 7. What value of y does the model predict when x = 33?
AnswerC — 172
Step-by-step- Substitute x = 33 into the line of best fit.
- ŷ = 5(33) + 7.
- ŷ = 165 + 7.
- So ŷ = 172.
Quick checkthe predicted point (33, 172) lies on the line ✓.
Common mistakes- A (5.2): solved for x instead of predicting y
- B (165): forgot the intercept
- D (236): swapped the slope and the intercept
Line Graphs
A line graph shows change over time; the steepest segment is the fastest change.
Key formulasLine graphs
Method. Percent change = (new − old) ÷ old × 100; always divide by the earlier year.
Worked exampleThe line graph shows concert tickets sold each year. What was the percent decrease from 2019 to 2021?

AnswerB — 80%
Step-by-step- Read the graph.
- So 2019 = 250 and 2021 = 50.
- The change is 200.
- Percent decrease = change ÷ original × 100 = 200 ÷ 250 × 100 = 80%.
Quick check250 − 80% of 250 = 250 − 200 = 50 ✓.
Common mistakes- A (20%): gave the new value as a percent of the original
- C (200%): gave the change, not the percent change
- D (400%): divided by the new value instead of the original
03.05 Probability and Conditional Probability
Problem-Solving and Data Analysis is about 15% of SAT Math (5–7 of 44 questions). This topic covers probability: favorable over total outcomes, the complement, probabilities from two-way tables, conditional probability (restrict to one row or column), independence, and expected counts.
Probability
Probability is favorable outcomes over equally likely total outcomes. For a conditional probability, restrict to the given row or column of a table. The expected count is the probability times the number of trials.
Key formulas| Basic | P(A)=totalfavorable |
| Complement | P(not A)=1−P(A) |
| Addition | P(A or B)=P(A)+P(B)−P(A and B) |
| Independent | P(A and B)=P(A)P(B) |
| Conditional | P(A∣B)=number in Bnumber in A and B |
| Expected count | n⋅P(A) |
Compound events
Method. Independent “and” → multiply the probabilities.
Worked exampleA coin and a die are used once each. What is the probability of getting heads from the coin and a 1 from the die?
AnswerA — 121
Step-by-step- The two results are independent.
- So multiply.
- P(heads) = 21 and P(a 1) = 61.
- So P(both) = 21 × 61 = 121.
Quick checkthere are 2 × 6 = 12 equally likely pairs and 1 is favorable, giving 121 ✓.
Common mistakes- B (61): used only the second event
- C (21): used only the first event
- D (32): added the probabilities
Addition rule
Method. Add the two probabilities, then subtract the overlap once.
Worked exampleFor two events, P(A) = 0.3, P(B) = 0.25 and P(A and B) = 0.2. What is P(A or B)?
AnswerB — 0.35
Step-by-step- Add P(A) and P(B).
- Then subtract the overlap P(A and B) once.
- So 0.3 + 0.25 − 0.2 = 0.35.
Quick check0.35 is at least the larger of P(A) and P(B) and at most 1 ✓.
Common mistakes- A (0.075): multiplied the probabilities
- C (0.55): did not subtract the overlap
- D (0.65): found the probability of neither
Expected counts
Method. Expected count = probability × number of spins.
Worked exampleA spinner has 8 green, 3 red and 4 purple equal sections. It is spun 75 times. About how many times should it land on green?
AnswerD — 40
Step-by-step- P(green) = 158.
- Multiply by the number of spins.
- So 158 × 75 = 40.
Quick checkthe expected counts for all three colors add to 75 ✓.
Common mistakes- A (158): gave the probability, not the count
- B (25): treated the three colors as equally likely
- C (35): found the expected number of non-green results
Conditional probability and two-way tables
Method. “Given” or “from those in” → divide by that group's total; “and” → one cell ÷ grand total.
Worked exampleThe table shows a survey of people. A person is chosen at random from those in the Prefers cats column. What is the probability that the person is in the Adults row?

A11039
B220107
C2213
D10778
AnswerD — 10778
Step-by-step- Only the Prefers cats column counts.
- So the total is 107.
- Of these, 78 are in the Adults row.
- So the probability is 10778.
Quick checkthe numerator and the denominator both come from the group the question names ✓.
Common mistakes- A (11039): divided by the grand total
- B (220107): found P(Prefers cats) instead
- C (2213): divided by the Adults row total
Conditional probability formula
Method. P(A | B) = P(A and B) ÷ P(B): divide by the probability of the given event.
Worked exampleP(A) = 0.8, P(B) = 0.3 and P(A and B) = 0.2. What is P(A | B), the probability of A given B?
AnswerD — 32
Step-by-step- P(A | B) means divide P(A and B) by P(B).
- So 0.2 ÷ 0.3 = 32.
Quick check32 × 0.3 = 0.2 = P(A and B) ✓.
Common mistakes- A (503): multiplied instead of dividing
- B (51): gave P(A and B)
- C (41): divided by P(A) instead of P(B)
Complement rule
Method. P(at least one) = 1 − P(none).
Worked exampleLena makes each free throw she attempts with probability 51, and her attempts are independent. If Lena attempts 3 free throws, what is the probability that she makes at least one of them?
A12561
B12564
C53
D125124
AnswerA — 12561
Step-by-step- “At least one” is the complement of “none”.
- P(none) = (1−51)3 = (54)3 = 12564.
- So P(at least one) = 1 − 12564 = 12561.
Quick checkP(none) + P(at least one) = 12564 + 12561 = 1 ✓.
Common mistakes- B (12564): found the probability that none succeed
- C (53): multiplied the probability by the number of tries
- D (125124): subtracted the all-succeed probability from 1
Two-Way Tables
Each cell counts members of a row group and a column group. For "given" questions divide by the row or column total, not the grand total.
Key formulas| Joint | grand totalcell |
| Conditional | row (or column) totalcell |
Two-way tables
Method. Of those who … → divide by that group's total.
Worked exampleThe two-way table shows the results of a survey of customers. Of the customers in the Tea column, what fraction are in the Evening row?

AnswerA — 14495
Step-by-step- The group is everyone in the Tea column.
- So divide by that column total, 144.
- Then 95 ÷ 144 = 14495.
Quick check14495 × 144 = 95, the Evening and Tea cell ✓.
Common mistakes- B (65): used the Coffee column
- C (2019): divided by the Evening row total
- D (2524): found the fraction of all customers in the Tea column
03.06 Inference from Samples and Margin of Error
Problem-Solving and Data Analysis is about 15% of SAT Math (5–7 of 44 questions). This topic covers estimating a population value from a random sample, and what a margin of error and a plausible interval mean. Larger random samples give smaller margins of error.
Inference and Margin of Error
A random sample's proportion or mean estimates the population value; multiply a sample proportion by the population size to estimate a count. The margin of error gives an interval of plausible values; larger samples give smaller margins of error.
Key formulas| Estimate a count | sample proportion×population size |
| Plausible interval | estimate±margin of error |
| Sample size | larger random sample⇒smaller margin of error |
Inference from a random sample
Method. Estimate = sample proportion × population size.
Worked exampleA random sample of 200 voters is taken from 400 voters in a district. Of the sample, 102 use public transit. About how many of all 400 voters use public transit?
AnswerD — 204
Step-by-step- In the sample, 200102 = 51% use public transit.
- A random sample represents the population.
- So estimate 200102 × 400 = 204.
Quick check400204 = 200102, the sample proportion ✓.
Common mistakes- A (51): scaled the wrong way
- B (102): gave the sample count
- C (196): estimated the number who do not
Margin of error
Method. Plausible values = estimate ± margin of error.
Worked exampleA poll of a random sample of students reports that 30% of the students favor a proposed schedule change, with a margin of error of ±2 percentage points. Based on these results, what is the smallest plausible value for the percent of all students who favor the change?
AnswerB — 28%
Step-by-step- The plausible values are the sample percent ± the margin of error, from 30 − 2 = 28% to 30 + 2 = 32%.
- So the answer is 28%.
Quick checkthe interval is centered at 30% and has width 2 × 2 = 4 points ✓.
Common mistakes- A (26%): doubled the margin of error
- C (30%): gave the sample percent
- D (32%): found the other end of the interval
03.07 Evaluating Statistical Claims
Problem-Solving and Data Analysis is about 15% of SAT Math (5–7 of 44 questions). This topic covers evaluating statistical claims: random sampling allows conclusions about the population sampled, random assignment allows cause-and-effect conclusions, and observational studies show only association. Watch for bias and for conclusions that go beyond the data.
Evaluating Statistical Claims
Only random sampling lets you generalize to the population that was sampled; only random assignment to treatments lets you conclude cause and effect. Observational studies show association only. Convenience and voluntary samples are biased.
Key formulas| Random sample | generalize to the sampled population only |
| Random assignment | cause-and-effect conclusion |
| Observational study | association, not causation |
Random sampling
Method. Random = chosen by chance from a list of everyone; not volunteers, not whoever is nearby.
Worked exampleA researcher wants a random sample of 40 employees from all employees at a company. Which method gives a random sample?
AAsk 40 employees chosen from the members of one club only.
BNumber every employee on a complete list and use a random number generator to choose 40 of them.
CPost an online survey and use the first 40 employees who choose to respond.
DAsk 40 employees who are friends of the person running the survey.
AnswerB — Number every employee on a complete list and use a random number generator to choose 40 of them.
Step-by-step- In a random sample every member of the population has the same chance of being chosen.
- Only numbering everyone on a complete list and choosing by a random number generator does that.
- The other methods let people choose themselves, pick whoever is handy, or leave out part of the population.
Quick checkthe chosen method gives every member of the population an equal chance of selection ✓.
Common mistakes- A (Ask 40 employees chosen from the members of one club only.): chose a sample from only one part of the population
- C (Post an online survey and use the first 40 employees who choose to respond.): chose a voluntary response sample
- D (Ask 40 employees who are friends of the person running the survey.): chose a sample of acquaintances, which is not random
Scope of conclusions
Method. Random sampling → generalize; random assignment → cause and effect.
Worked exampleResearchers randomly selected 200 voters from all voters in a district, recorded which exercise routine each already follows, and measured each voter's resting heart rate. They found that resting heart rate was related to the exercise routine followed. Which of the following conclusions about this relationship is justified?
AThe result can be generalized to all voters in the district, but cause and effect cannot be concluded.
BCause and effect can be concluded, and the result can be generalized to all voters in the district.
CCause and effect can be concluded only for the subjects in the study.
DNeither cause and effect nor generalization to all voters in the district is justified.
AnswerA — The result can be generalized to all voters in the district, but cause and effect cannot be concluded.
Step-by-step- Random sampling allows generalizing to the population that was sampled, all voters in the district.
- Random assignment allows cause-and-effect conclusions.
- Here the voters were randomly selected, but each voter chose his or her own routine (no random assignment).
- So the result can be generalized to all voters in the district, but cause and effect cannot be concluded.
Quick checkrandom sampling is present and random assignment is not present, matching the chosen conclusion ✓.
Common mistakes- B (Cause and effect can be concluded, and the result can be generalized to all voters in the district.): thought there was random assignment
- C (Cause and effect can be concluded only for the subjects in the study.): missed the random sampling and thought there was random assignment
- D (Neither cause and effect nor generalization to all voters in the district is justified.): missed the random sampling