Skip to content
SAT essentials · Math

SAT probability and two-way tables

Every SAT probability is a part over a whole. The skill is choosing the right whole, especially when a table or a condition shrinks the group.

These are the concepts this skill rests on, the same pages as in the lesson.

Chance as a Part Over a Whole

A Chance Is a Part Over a Whole

"Selected at random" means every outcome is equally likely. Then the probability of an event is the part over the whole:

A box of 30 tiles drawn as 30 circles in three rows of ten. Twelve are filled green and labeled the part; all 30 are labeled the whole. The probability of drawing a green tile is 12 over 30, which is 2 over 5. A box of 30 tiles, each equally likely to be drawn part: 12 green whole: all 30 tiles P(green) = 12/30 = 2/5

P(event)=event outcomesall outcomesP(\text{event}) = \frac{\text{event outcomes}}{\text{all outcomes}}

A box holds 30 tiles, and 12 are green. Draw one at random: P(green)=1230=25P(\text{green}) = \frac{12}{30} = \frac{2}{5}.

A probability is always between 0 and 1. A 0 means the event can't happen; a 1 means it's certain. If you get more than 1, something went wrong: usually the part and whole got swapped, or an overlap got counted twice. Write the whole first, then the part: most wrong answers in this skill divide by the wrong whole.

Find the Whole First

A one-way table lists a count for each category. If it has no total row, add up every category before you divide.

A bar split into three parts for a playlist: 14 rock songs, 9 jazz songs shaded orange, and 7 folk songs. A bracket under the whole bar reads whole: 14 + 9 + 7 = 30 songs. Below: P(jazz) = 9/30 = 3/10, not 1/3. A playlist of rock, jazz and folk songs rock: 14 jazz: 9 folk: 7 whole: 14 + 9 + 7 = 30 songs P(jazz) = 9/30 = 3/10 not 1/3: the three kinds are not equally common

A playlist has 14 rock, 9 jazz and 7 folk songs. The whole is 14+9+7=3014 + 9 + 7 = 30, so P(jazz)=930=310P(\text{jazz}) = \frac{9}{30} = \frac{3}{10}. Three kinds of song doesn't make each kind 13\frac{1}{3} likely: each song has the same chance, not each kind.

A count can also be given as an expression. If a box holds xx red pens and 3 times as many blue pens, and nothing else, the whole is x+3x=4xx + 3x = 4x, and P(blue)=3x4x=34P(\text{blue}) = \frac{3x}{4x} = \frac{3}{4}.

Fractions, Decimals, and Percents

You can write the same chance three ways: 38\frac{3}{8}, 0.375 and 37.5% are the same number.

A bar from 0 to 1 cut into 8 equal parts, with 3 parts shaded blue. At the end of the shaded part: 3/8 = 0.375 = 37.5%. One chance, three ways to write it 0 1 3 of 8 parts 3/8 3/8 = 0.375 = 37.5%
  • If the question gives a percent of a group, that's already a chance: if 25% of the members are new, P(new)=14P(\text{new}) = \frac{1}{4}.
  • A chance asked for as a percent is the fraction times 100: 720\frac{7}{20} is 35%.
  • Convert once, at the end. Multiply by 100 too early and the chance 0.35 turns into 35, which isn't a probability.

Typing an answer. The answer field takes up to five characters, six with a minus sign. A fraction is safest, and you don't need to reduce it: 6/16 scores the same as 3/8. The leading zero is optional, so .375 and 0.375 both work. A decimal that keeps going must fill the whole field: for 23\frac{2}{3}, .6666 or .6667 scores, and .67 does not.

Counting Whole Numbers in a Range

Some questions give you no counts at all: "a whole number from 5 to 20, inclusive, is selected at random." Then you count the whole and the part yourself.

A row of boxes for the whole numbers 5 through 20. The multiples of 3, which are 6, 9, 12, 15 and 18, are shaded green. Below: count 20 minus 5 plus 1 = 16 numbers; 5 multiples of 3; so the probability is 5/16. Whole numbers from 5 to 20, inclusive 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 whole: 20 − 5 + 1 = 16 numbers part: 6, 9, 12, 15, 18 is 5 numbers P(multiple of 3) = 5/16
  1. Count the whole. "Inclusive" keeps both ends: 20−5+1=1620 - 5 + 1 = 16 numbers, not 15.
  2. Find the first and last numbers that qualify. For multiples of 3, they are 6 and 18.
  3. Count them by steps: 18−63+1=5\frac{18 - 6}{3} + 1 = 5. With small numbers, listing them works too: 6, 9, 12, 15, 18.
  4. Divide: P(multiple of 3)=516P(\text{multiple of 3}) = \frac{5}{16}.

Guessing "one in every three" gives 13\frac{1}{3}, which is wrong here: 16 numbers don't split evenly into groups of three.

Combining Events

Not: The Complement

Every outcome is either in event AA or not in it, never both. So the two chances add to 1:

A bar split into two parts that fill it: late, 0.3, shaded orange, and not late, 0.7, shaded blue. Below: 0.3 + 0.7 = 1, so P(not late) = 1 − 0.3. Every bus is late or not late late: 0.3 not late: 0.7 0.3 + 0.7 = 1 P(not late) = 1 − 0.3 = 0.7

P(not A)=1−P(A)P(\text{not } A) = 1 - P(A)

If a randomly chosen bus is late with probability 0.3, it is not late with probability 1−0.3=0.71 - 0.3 = 0.7.

With counts, subtract the event's count from the whole. In a box of 40 pens where 12 are red, 28 are not red, and P(not red)=2840=710P(\text{not red}) = \frac{28}{40} = \frac{7}{10}. When the event covers several categories but its opposite covers just one, the complement is the shortcut.

Name the opposite in words before you subtract. The opposite of "at least one" is "none," not "exactly one."

Or: Subtract the Overlap Once

"AA or BB" means AA, BB, or both. Adding the two groups counts the members of both twice, so take the overlap out once:

Two overlapping circles inside a box of 40 students. The left circle, choir: 18, holds 13 alone; the right circle, band: 12, holds 7 alone; the overlap, shaded orange, holds 5; 15 students are outside both. Below: choir or band = 18 + 12 − 5 = 25, because adding 18 and 12 counts the 5 twice. choir: 18 band: 12 13 5 7 neither: 15 40 students choir or band: 18 + 12 − 5 = 25 18 + 12 counts the 5 in both twice

P(A or B)=P(A)+P(B)−P(both)\begin{aligned} P(A \text{ or } B) &= P(A) + P(B) \\ &\quad - P(\text{both}) \end{aligned}

Of 40 students, 18 sing in the choir, 12 play in the band, and 5 do both. Choir or band: 18+12−5=2518 + 12 - 5 = 25 students, so the probability is 2540=58\frac{25}{40} = \frac{5}{8}.

When two events cannot both happen, the overlap is 0 and you just add. "Each student has exactly one homeroom" means no student is in two, so P(room 1 or room 2)P(\text{room 1 or room 2}) is the sum of the two chances. Read the situation to tell which case you're in. If the two groups add up to more than the whole, they must overlap.

In a two-way table, "or" across a row and a column is the row total plus the column total minus the cell they share. It is also the grand total minus the cells that are neither.

Exactly One, and Neither

Exactly one means one of the two but not both. Take the overlap out of each group:

The same two circles for 40 students. The parts of each circle outside the overlap, 13 and 7, are shaded green; the overlap of 5 is gray. Below: exactly one of the two is 13 + 7 = 20 students; neither is 40 − 25 = 15. choir: 18 band: 12 13 5 7 neither: 15 40 students exactly one: 13 + 7 = 20 neither: 40 − 25 = 15

With the choir and band above: choir only 18−5=1318 - 5 = 13, band only 12−5=712 - 5 = 7. Exactly one is 13+7=2013 + 7 = 20 students, so the probability is 2040=12\frac{20}{40} = \frac{1}{2}. The same number comes from "or" minus "both": 25−5=2025 - 5 = 20.

Neither is everyone outside both groups: 40−25=1540 - 25 = 15 students, so P(neither)=1540=38P(\text{neither}) = \frac{15}{40} = \frac{3}{8}.

You can run these backward, too. If you're told how many are in neither, then "or" is the whole minus neither, and "both" is A+B−(or)A + B - (\text{or}).

Reading Two-Way Tables

Reading a Two-Way Table

A two-way table sorts one group by two things at once. Each inside cell counts the members with one value of each. The margins total a row or a column, and the corner is the grand total, where every member is counted once.

A two-way table of 120 students by grade and how they get to school. The cell for 9th grade and walks, 24, is shaded blue; the 9th-grade row total, 60, is shaded orange; the grand total, 120, is shaded green. Walks Bus Total 9th 24 36 60 10th 16 44 60 Total 40 80 120 a cell: 9th grade and walks, 24 a margin: every 9th grader, 60 the grand total: every student, 120
  • Two conditions joined by "and" name one cell. A 9th grader who walks: P=24120=15P = \frac{24}{120} = \frac{1}{5}.
  • One condition names a margin. A 9th grader: P=60120=12P = \frac{60}{120} = \frac{1}{2}.
  • With no restriction, the whole is the grand total.

Check that the table adds up before you use it: 24+36=6024 + 36 = 60 across, and 24+16=4024 + 16 = 40 down.

Recover a Missing Cell

A blank cell is a question, not a zero. Fill it in from its row or column before you find any probability.

A two-way table with one blank cell, shaded orange, in row A under column Yes. Row A has 32 in No and a total of 50, so the blank is 50 − 32 = 18. Check down the column: 18 + 22 = 40, the column total. No Yes Total Group A 32 ? 50 Group B 28 22 50 Total 60 40 100 across its row: ? = 50 − 32 = 18 check down its column: 18 + 22 = 40

The blank is its row total minus the other cell in its row: 50−32=1850 - 32 = 18. Then check it down its column: 18+22=4018 + 22 = 40.

  • A blank margin is the sum of its row or column, or the grand total minus the other margin.
  • With two blanks, fill the one you can first; it gives you the other.

When the Group Shrinks

The Words That Shrink the Group

A conditional probability is a chance inside a smaller group. The question names the group, and that group's total becomes the whole:

A two-way table of 100 students by class time and whether they passed. The evening row is shaded blue and labeled the new whole, 40 students; its passed cell, 26, is the part. The morning row and the totals row are grayed out. Passed Did not Total Morning 45 15 60 Evening 26 14 40 Total 71 29 100 new whole: the 40 evening students part: the 26 of them who passed

P(A given B)=number in bothnumber in BP(A \text{ given } B) = \frac{\text{number in both}}{\text{number in } B}

Questions name the group in lots of ways: given that, of the, among, of those who, if the selected student is, consider only. "Of the evening students, what is the probability that a randomly selected one passed?" The whole is the 40 evening students and the part is the 26 who passed: 2640=1320\frac{26}{40} = \frac{13}{20}.

Dividing 26 by the grand total, 100, answers a different question: the chance that a student picked from everyone is an evening student who passed.

Row or Column: Two Different Questions

One cell can sit over two different wholes. Which one you use depends on which group is given.

The same two-way table drawn twice. Top: the Sport row is shaded, a whole of 40, so given sport, band is 12 of 40. Bottom: the Band column is shaded, a whole of 30, so given band, sport is 12 of 30. The cell 12 is the part both times. Band No band Total Sport 12 28 40 No sport 18 22 40 Total 30 50 80 given sport: 12 of 40 Band No band Total Sport 12 28 40 No sport 18 22 40 Total 30 50 80 given band: 12 of 30
  • Given sport, the chance of band is 1240=310\frac{12}{40} = \frac{3}{10}: the whole is the Sport row.
  • Given band, the chance of sport is 1230=25\frac{12}{30} = \frac{2}{5}: the whole is the Band column.

Find the given group first: its total goes on the bottom. The other condition picks the cell inside it. Swap the two and you answer the reverse question, which is a favorite wrong choice.

A Condition That Spans Rows

"At least 20 years old," "not under 20," and "under 40" each cover more than one row. Add those rows' totals for the whole, and the matching cells for the part.

A two-way table with three age rows. The rows 20 to 39 and 40 or older are shaded blue: together they are the members at least 20 years old. Below: whole 50 + 30 = 80, part 21 + 9 = 30. Age Yes No Total Under 20 12 18 30 20 to 39 21 29 50 40 or older 9 21 30 Total 42 68 110 whole, at least 20: 50 + 30 = 80 part, also yes: 21 + 9 = 30

Of the members at least 20 years old, the chance of answering yes is

21+950+30=3080=38\frac{21 + 9}{50 + 30} = \frac{30}{80} = \frac{3}{8}

"Not under 20" is the same group, and its total is also the grand total minus one row: 110−30=80110 - 30 = 80.

Build the Table From Words

When the counts come in sentences, draw the table yourself. Fill in the numbers you're given, complete the rest by subtracting and adding, then read it like any other table.

A two-way table for 120 visitors by age group and whether they bought a map. The numbers stated in words, 50 adults, 20 adults with a map, 28 children with a map and 120 in all, are blue. The rest, found by subtracting or adding, are green: 70 children, 30 and 42 without a map, and 48 and 72 in the column totals. Map No map Total Adult 20 30 50 Child 28 42 70 Total 48 72 120 blue: stated in the words green: found by subtracting or adding Every row and column must add up.

"A museum had 120 visitors; 50 were adults. Of the adults, 20 bought a map, and of the children, 28 bought a map." Children: 120−50=70120 - 50 = 70. Map buyers: 20+28=4820 + 28 = 48.

  • Of the visitors who bought a map, P(adult)=2048=512P(\text{adult}) = \frac{20}{48} = \frac{5}{12}.
  • Of the adults, P(map)=2050=25P(\text{map}) = \frac{20}{50} = \frac{2}{5}, a different question.

A percent becomes a count first: "35% of 400 employees" is 140 employees.

Compare Rates, Not Counts

"Which group is more likely to…" compares two conditional probabilities. Divide each group's count by its own total, then compare.

Two bars of the same length, each standing for its own whole. Store A: 66 of 110 orders on time, 0.6 of its bar shaded. Store B: 34 of 40 orders on time, 0.85 of its bar shaded. A has the larger count; B has the larger rate. Each bar is one store's own whole Store A: 66 of its 110 orders on time 0.6 Store B: 34 of its 40 orders on time 0.85 A has more on time; B has the higher rate

Store A delivered 66 of its 110 orders on time: 66110=0.6\frac{66}{110} = 0.6. Store B delivered 34 of its 40: 3440=0.85\frac{34}{40} = 0.85. B's orders were more likely to be on time, even though A had more on-time orders. A just had more orders.

A question can also compare one rate with a target: "Is a randomly chosen order from Store A more likely than not to be on time?" Yes, since 0.6>120.6 > \frac{1}{2}.

Independent Events

Two events are independent when knowing that one happened does not change the chance of the other. Then, and only then, the chance that both happen is the product:

A tree diagram for two machines that work or fail independently. Machine A works with probability 0.6 and fails with 0.4; machine B works with 0.9 and fails with 0.1, whichever way A went. Each path multiplies: works, works is 0.6 × 0.9 = 0.54, shown in green; the others are 0.06, 0.36 and 0.04, and the four add to 1. machine A machine B 0.6 works 0.4 fails 0.9 works 0.6 × 0.9 = 0.54 0.1 fails 0.6 × 0.1 = 0.06 0.9 works 0.4 × 0.9 = 0.36 0.1 fails 0.4 × 0.1 = 0.04

P(A and B)=P(A)⋅P(B)P(A \text{ and } B) = P(A) \cdot P(B)

Two machines work independently, with probabilities 0.6 and 0.9. Both work with probability 0.6×0.9=0.540.6 \times 0.9 = 0.54. In the tree, follow a path and multiply along it; the four paths add to 1.

From a table, test it: compare a row's rate with the overall rate. If 7 of the 20 members of a row have a trait and 28 of all 80 members do, both rates are 0.35, and the two events are independent. Rates of 0.35 and 0.36 are close but not equal, so those events are not independent.

Independent is not the same as "cannot both happen." Two events that each can happen but cannot happen together are not independent: knowing one happened tells you the other did not.

At Least One

"At least one" is every outcome except the ones where none happen, so take the chance of none away from 1. In the tree, the orange path is the only one where neither machine works; the three green paths are all "at least one."

The same tree for two independent machines. The three paths where at least one machine works, 0.54, 0.06 and 0.36, end in green. The one path where neither works, fails then fails, is orange along both branches: 0.4 × 0.1 = 0.04. Below: at least one works is 1 − 0.04 = 0.96, the same as 0.54 + 0.06 + 0.36. machine A machine B 0.6 works 0.4 fails 0.9 works 0.54 0.1 fails 0.06 0.9 works 0.36 0.1 fails 0.04 neither works: 0.04 at least one: 1 − 0.04 = 0.96

P(at least one)=1−P(none)P(\text{at least one}) = 1 - P(\text{none})

For the two independent machines, neither works with probability 0.4×0.1=0.040.4 \times 0.1 = 0.04, so at least one works with probability 1−0.04=0.961 - 0.04 = 0.96. The long way adds up three paths: 0.54+0.06+0.36=0.960.54 + 0.06 + 0.36 = 0.96.

Exactly one is the two mixed paths: 0.06+0.36=0.420.06 + 0.36 = 0.42.

Counts From Chances

Expected Count

A probability predicts a count. The expected number in a group is

expected count=chance×group size\begin{aligned} &\text{expected count} \\ &\quad = \text{chance} \times \text{group size} \end{aligned}

Forty circles in four rows of ten, ten of them filled blue: a sample in which 10 of 40, or 1/4, have the trait. Below: expected in a group of 400: 1/4 × 400 = 100. A sample of 40: 10 have the trait rate: 10/40 = 1/4 in 400: 1/4 × 400 = 100

In a sample, 10 of 40 have a trait, a rate of 14\frac{1}{4}. In a group of 400, expect 14×400=100\frac{1}{4} \times 400 = 100. The answer is a count, not a chance.

  • Use the rate of the group you are predicting about. A prediction about adults uses the adult row's rate, not the rate for everyone.
  • When the product is not a whole number, round as the question says, for example "to the nearest whole number."

Work Backward to a Count

When you know the chance but a count is missing, set part over whole equal to the chance, put the unknown where it belongs, and solve.

A bag holds 6 red marbles and bb blue marbles, and P(red)=25P(\text{red}) = \frac{2}{5}. Think of the bag as 5 equal shares: red makes up 2 shares. The 6 red marbles make 2 shares, so each share has 3 marbles, and the 3 blue shares have 9. As an equation:

66+b=25\frac{6}{6 + b} = \frac{2}{5}

6 is 25\frac{2}{5} of the whole, so the whole, 6+b6 + b, is 15, and b=9b = 9. Check: 615=25\frac{6}{15} = \frac{2}{5}.

Adding blue marbles raises only the whole, so the chance of red falls. In the widget below, the whole rectangle always represents 100% of the chance, not the physical size of the bag. The red count stays 6; its share of the area shrinks as the blue count grows.

Start with no blue marbles: red fills the whole rectangle. Add blue until red reaches the dashed 40% mark. At 9 blue marbles, red is 615=25=40%\frac{6}{15} = \frac{2}{5} = 40\%, exactly the chance you want.

A bar for a bag of red and blue marbles, cut into 5 equal parts because P(red) = 2/5. Two parts are red, bracketed 6 red; three parts are blue, bracketed b blue. Each part is labeled 3. Below: 6 ÷ 2 = 3 marbles in each part, so the bag holds 5 × 3 = 15 and b = 3 × 3 = 9. 3 3 3 3 3 6 red b blue P(red) = 2/5: 2 of 5 equal parts 6 red ÷ 2 parts = 3 in each part whole: 5 × 3 = 15 marbles b = 3 parts × 3 = 9 blue

Keep going

Check Your Skills, the strategy for each kind of question, worked examples and practice from Easy to Hard are in the Quaero SAT Course.

More Problem-Solving and Data Analysis