AP Statistics · Unit 2 · Probability, Random Variables and Probability Distributions
2.2 Two categorical variables: summary statistics
Calculate marginal and conditional distributions from a two-way table, compare them using percentages, and draw conclusions about association.
1/5A marginal distribution describes one variable by itself
- A marginal distribution describes one variable by itself
- A conditional distribution describes one variable within a group
- Percentages within groups allow fair comparisons
- Different conditional distributions indicate an association
- Association in a two-way table does not prove cause and effect
- Check your recall
- Practice: 10 questions
Part 1 of 5: A marginal distribution describes one variable by itself
A survey of 300 adults records each person’s age group and whether they use a fitness tracker. The counts go into a two-way table. We can describe either variable alone or both together.
Tracker use by age group (number of adults)
- Age group
- Under 30
- Uses tracker
- 75
- No tracker
- 25
- Total
- 100
- Age group
- 30–49
- Uses tracker
- 60
- No tracker
- 60
- Total
- 120
- Age group
- 50 and over
- Uses tracker
- 30
- No tracker
- 50
- Total
- 80
- Age group
- Total
- Uses tracker
- 165
- No tracker
- 135
- Total
- 300
A marginal distribution gives the percent of all units in each category of one variable, ignoring the other. For tracker use, divide each column total by the 300 adults: use one, and do not. The age groups make up 33.3%, 40.0% and 26.7% of the sample. These percents describe the whole sample at once.
Key term
marginal distribution
A marginal distribution is the distribution of one variable in a two-way table, found by dividing its totals by the total of all units.
Part 2 of 5: A conditional distribution describes one variable within a group
Marginal percents ignore age, so they cannot show whether age matters. A conditional distribution describes one variable among only the units in one category of the other. It gives a percent for each category. The denominator is the group’s own total. It is not the total for everyone.
Worked example
Finding a conditional distribution
- Word equation: conditional percent = cell count divided by group total.
- The group total for adults under 30 is 100.
- Uses a tracker: .
- No tracker: .
Answer: Among adults under 30, 75% use a tracker and 25% do not.
Each group has its own conditional distribution, and each one adds to 100%.
Key term
conditional distribution
A conditional distribution is the distribution of one variable among only the units in one category of the other variable.

Quick check
One question on the part you just read
Easy
Show answer and explanation
Answer: B · Worked explanation
Part 3 of 5: Percentages within groups allow fair comparisons
The groups here have 100, 120 and 80 adults, so raw counts cannot be compared directly. Calculate the percents within each group, using that group’s total. The groups should be the ones named in the question. To ask whether tracker use depends on age, find percents within each age group.
Percent using a tracker within each age group
- Age group
- Under 30
- Uses tracker
- 75.0%
- No tracker
- 25.0%
- Age group
- 30–49
- Uses tracker
- 50.0%
- No tracker
- 50.0%
- Age group
- 50 and over
- Uses tracker
- 37.5%
- No tracker
- 62.5%
Reading down the table, the percent using a tracker falls as age rises. Here, the gap between 75.0% and 37.5% is 37.5 units, each called a percentage point. Use this unit for a difference between two percents.
Key term
percentage point
A percentage point is the unit used for the difference between two percents, such as 75% and 50%.
Part 4 of 5: Different conditional distributions indicate an association
Comparing the three rows shows how the pattern changes from one group to the next. If the conditional distributions differ noticeably, the variables have an association. Here the tracker percents fall steadily from 75.0% to 37.5% with age. Age group and tracker use are associated in this sample.
If the conditional distributions were nearly identical, the table would show little or no association. For example, 50% in every age group would mean that age tells us nothing about tracker use.
Key term
association
An association exists between two variables when the distribution of one variable differs across the categories of the other.

Quick check
One question on the part you just read
Easy
Show answer and explanation
Answer: C · Worked explanation
Part 5 of 5: Association in a two-way table does not prove cause and effect
A clear association invites an explanation, but the table alone cannot supply one. Data from an observational study, in which researchers assign no treatments, can show that two variables move together. They cannot show that one causes the other.
Another factor linked to both variables might explain the pattern. Older adults may differ in income or comfort with technology. Cause-and-effect conclusions need a randomized experiment, so report only the pattern in the data.
Chance also matters. With few units, a small difference in percents can appear without any real relationship.
Key term
observational study
An observational study is a study in which researchers measure variables without assigning treatments to the units.
Check your recall[12 marks]
5 written questions. Answer each one in your head or on paper, then open the mark scheme.
- 1.Describe[3 marks]
Describe the difference between a marginal distribution and a conditional distribution.
Show mark schemeHide mark scheme
Model answer
A marginal distribution describes one variable using all units and ignores the other variable. A conditional distribution describes one variable within a single category of the other. Marginal percents divide by the total of all units, and conditional percents divide by that group’s total.
- 1 mark for stating that a marginal distribution describes one variable for all units, ignoring the other
- 1 mark for stating that a conditional distribution describes one variable within one category of the other
- 1 mark for noting that the denominators differ: all units for marginal and the group total for conditional
- 2.Calculate[3 marks]
In a survey of 400 people, 160 live in cities, and 96 of those city residents prefer online shopping. In all, 180 of the 400 people prefer online shopping. Calculate the percent of city residents who prefer online shopping and the marginal percent who prefer online shopping.
Show mark schemeHide mark scheme
Model answer
City residents: conditional percent = , so 60%. All people: marginal percent = , so 45%.
- 1 mark for dividing 96 by the city total 160
- 1 mark for obtaining 60%
- 1 mark for dividing 180 by 400 to obtain 45%
- 3.Justify[3 marks]
Of 300 customers who used coupons, 45% bought a second item. Of 200 customers who did not use coupons, 45% bought a second item. Justify whether coupon use and buying a second item show an association.
Show mark schemeHide mark scheme
Model answer
The conditional distributions are the same for both groups, with 45% buying a second item in each. The distribution of the response does not differ between groups, so there is little or no association.
- 1 mark for comparing the conditional percents for the two groups
- 1 mark for stating that the percents are equal, so the distributions do not differ
- 1 mark for concluding that there is little or no association
- 4.Explain[2 marks]
Explain why a strong association in an observational study does not show that one variable causes the other.
Show mark schemeHide mark scheme
Model answer
Researchers assigned no treatments, so another factor linked to both variables could explain the pattern. Only a randomized experiment supports cause and effect.
- 1 mark for stating that another factor linked to both variables could explain the pattern
- 1 mark for stating that only a randomized experiment supports a cause-and-effect conclusion
- 5.Identify[1 mark]
Identify the number to divide by to find the percent of adults under 30 who use a tracker, given that 75 of the 100 adults under 30 use one and 300 adults were surveyed.
Show mark schemeHide mark scheme
Model answer
Divide by 100, the total for the under-30 group.
- 1 mark for identifying 100, the group total

Practice this topic
10 quick-fire multiple-choice questions, easy to hard. Tap an answer to see if you're right and the worked explanation. (The written questions with mark schemes are under “Check your recall” above.)
0 right0 of 10 answered
Includes the 2 quick checks in the note above.
EasyIdentify[1 mark]
Show answer and explanation
Answer: A · Worked explanation
MediumCalculate[2 marks]
Show answer and explanation
Answer: D · Worked explanation
MediumCalculate[2 marks]
Show answer and explanation
Answer: B · Worked explanation
MediumDetermine[2 marks]
Show answer and explanation
Answer: A · Worked explanation
MediumInterpret[2 marks]
Show answer and explanation
Answer: D · Worked explanation
HardDetermine[2 marks]
Show answer and explanation
Answer: C · Worked explanation
HardDetermine[2 marks]
Show answer and explanation
Answer: B · Worked explanation
HardDetermine[2 marks]
Show answer and explanation
Answer: A · Worked explanation