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 holds 30 tiles, and 12 are green. Draw one at random: .
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 playlist has 14 rock, 9 jazz and 7 folk songs. The whole is , so . Three kinds of song doesn't make each kind likely: each song has the same chance, not each kind.
A count can also be given as an expression. If a box holds red pens and 3 times as many blue pens, and nothing else, the whole is , and .
Fractions, Decimals, and Percents
You can write the same chance three ways: , 0.375 and 37.5% are the same number.
- If the question gives a percent of a group, that's already a chance: if 25% of the members are new, .
- A chance asked for as a percent is the fraction times 100: 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 , .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.
- Count the whole. "Inclusive" keeps both ends: numbers, not 15.
- Find the first and last numbers that qualify. For multiples of 3, they are 6 and 18.
- Count them by steps: . With small numbers, listing them works too: 6, 9, 12, 15, 18.
- Divide: .
Guessing "one in every three" gives , 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 or not in it, never both. So the two chances add to 1:
If a randomly chosen bus is late with probability 0.3, it is not late with probability .
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 . 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
" or " means , , or both. Adding the two groups counts the members of both twice, so take the overlap out once:
Of 40 students, 18 sing in the choir, 12 play in the band, and 5 do both. Choir or band: students, so the probability is .
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 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:
With the choir and band above: choir only , band only . Exactly one is students, so the probability is . The same number comes from "or" minus "both": .
Neither is everyone outside both groups: students, so .
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 .
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.
- Two conditions joined by "and" name one cell. A 9th grader who walks: .
- One condition names a margin. A 9th grader: .
- With no restriction, the whole is the grand total.
Check that the table adds up before you use it: across, and 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.
The blank is its row total minus the other cell in its row: . Then check it down its column: .
- 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:
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: .
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.
- Given sport, the chance of band is : the whole is the Sport row.
- Given band, the chance of sport is : 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.
Of the members at least 20 years old, the chance of answering yes is
"Not under 20" is the same group, and its total is also the grand total minus one row: .
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 museum had 120 visitors; 50 were adults. Of the adults, 20 bought a map, and of the children, 28 bought a map." Children: . Map buyers: .
- Of the visitors who bought a map, .
- Of the adults, , 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.
Store A delivered 66 of its 110 orders on time: . Store B delivered 34 of its 40: . 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 .
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:
Two machines work independently, with probabilities 0.6 and 0.9. Both work with probability . 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."
For the two independent machines, neither works with probability , so at least one works with probability . The long way adds up three paths: .
Exactly one is the two mixed paths: .
Counts From Chances
Expected Count
A probability predicts a count. The expected number in a group is
In a sample, 10 of 40 have a trait, a rate of . In a group of 400, expect . 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 blue marbles, and . 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:
6 is of the whole, so the whole, , is 15, and . Check: .
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 , exactly the chance you want.
Keep going
Build the basics first
Foundations has free readiness checks and lessons in the English and math the SAT expects.
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
Ratios, rates and units
Set up ratios and proportions, work with rates, convert units by cancelling them, and recognize proportional relationships.
Read the concepts →SAT essentials · MathPercentages
Learn what a percent means, find shares of a total, and apply one percent change or several in a row.
Read the concepts →SAT essentials · MathOne-variable data
Find the mean and median, read dot plots, histograms and box plots, compare spread, and see how one value changes a data set.
Read the concepts →SAT essentials · MathTwo-variable data
Read two-variable data, predict with a line of best fit, understand residuals, interpret slope and intercept in context, and choose a model.
Read the concepts →SAT essentials · MathSample statistics and margin of error
Tell a sample from a population, build an interval from an estimate and its margin of error, see why sample size matters, and scale results up.
Read the concepts →SAT essentials · MathEvaluating statistical claims
Tell observational studies from experiments, and see what random selection and random assignment each allow a study to conclude.
Read the concepts →