find row and column relative frequencies in a two-way table and use them to describe a possible association between two categorical variables.
Phase 4: Build
Do Now. Read the CDC’s sleep advice. Then decide: in the phone and sleep sample, 18 phone students and 24 no-phone students got 8 or more hours. Does that mean the no-phone group did only a little better?
The CDC page recommends how much sleep for teens ages 13 to 17?
The reading opens over the lesson; close it to come back to exactly where you were.
Counts can fool us. In the sample, 18 phone students and 24 no-phone students got 8 or more hours. But the groups are different sizes: 60 and 40. To compare fairly, divide each cell by its group total. That gives a relative frequency, which we can write as a fraction, a decimal or a percent.
| Sample survey, 100 eighth graders | 8 or more hours of sleep | Fewer than 8 hours | Total |
|---|---|---|---|
| Phone in the bedroom at night | 18 | 42 | 60 |
| No phone in the bedroom | 24 | 16 | 40 |
| Total | 42 | 58 | 100 |
Worked example 1: Zoe uses row relative frequencies. Zoe’s question is: does the share of students getting enough sleep differ for the two phone groups? She divides by row totals. Phone: 18 ÷ 60 = 0.30 = 30%. No phone: 24 ÷ 40 = 0.60 = 60%. The no-phone group’s rate is twice as high. Because the percents are very different, the sample suggests an association between keeping a phone in the bedroom and getting less sleep. Each row’s percents add to 100%: 30% + 70% and 60% + 40%.
Worked example 2: Eli finds no association. In another sample, 20 students are in a club and 12 of them read for fun each week. Of 30 not in a club, 18 read for fun. Club: 12 ÷ 20 = 60%. No club: 18 ÷ 30 = 60%. The rates are the same, so this sample shows no association between club membership and reading for fun, even though 18 is more than 12.
Rows or columns? Let the question decide. “Of the students who got 8 or more hours, what share had a phone?” is a column question: 18 ÷ 42 ≈ 0.429, or about 43%. Always say which total you divided by.
Real data too. In the ERS table from 1.20, households with children: 6.7 ÷ 36.6 ≈ 18.3% food insecure. Households without children: 11.6 ÷ 97.4 ≈ 11.9%. That suggests an association between having children at home and food insecurity. (The ERS page reports 18.4%, computed from exact counts; our rounded millions give 18.3%.)
Association is not cause. Our phone table cannot show that phones cause less sleep. Students with phones might also have later activities, or something else might explain both.
Write each relative frequency as a percent. Say which total you used.
Finish your panel’s two-way table finding.
For your team's two-way table (or the sport and breakfast sample), find the row relative frequencies, write them in the table in a second color, and decide whether the data suggest an association.
Use the reading. Explain why our survey uses 8 hours as its cut-off. Would a survey of fifth graders need a different cut-off? Use the CDC table.
Compare rates, not raw counts, when groups are different sizes.
Exit ticket: explain why 24 is less than twice 18, but 60% is twice 30%.
Pacing: warm-up and reading 7 minutes, mini lecture 10, guided practice 15, independent 6, closure 2.
Likely misconceptions: comparing raw counts across groups of different sizes; dividing by the grand total when the question is about a row; concluding cause from an association.
Supports: a highlighting routine: circle the total you divide by before computing. Stretch: change one count in the phone table so the association disappears, and explain your change.
Note: the CDC table is from the About Sleep page (opened 2026-11-20). Our survey asks about school nights only; the 8-hour cut-off is the low end of the CDC teen range.
Answer guide: warm-up 8 to 10 hours; 30%; 60; association 75% vs 50%; false; about 43%; 12 (11.9%); both rates are 60%. Board finding: “In our sample, 30% of the 60 students with a phone in the bedroom got 8 or more hours of sleep, compared with 60% of the 40 students without one. This suggests an association between phones in the bedroom and less sleep, but it does not show that phones cause it.” Fifth graders: the CDC range for ages 6-12 is 9 to 12 hours, so a 9-hour cut-off fits. Exit ticket: 24 ÷ 18 = 1.33, but 24 comes from a group of 40 and 18 from a group of 60, and 0.60 ÷ 0.30 = 2.
Centers for Disease Control and Prevention, About Sleep · U.S. government work, public domain
The Centers for Disease Control and Prevention, or CDC, is the nation’s public health agency. Its About Sleep page gives a table of how much sleep people need at each age.
Needs change as we grow. The table recommends 9 to 12 hours a day for school-age children ages 6 to 12, and 8 to 10 hours for teens ages 13 to 17. For adults ages 18 to 60, it recommends 7 or more hours. Babies and toddlers need even more, including naps.
The page explains that getting enough sleep is linked to good health. It can help people get sick less often, and it is linked to a lower risk of serious conditions such as type 2 diabetes, heart disease, high blood pressure and stroke. Enough sleep can also lower the risk of car crashes and the injuries they cause.
Quality matters, not just hours. The page lists signs of poor sleep, such as trouble falling asleep, waking up again and again during the night, and feeling sleepy or tired even after enough hours in bed.
For our survey, most eighth graders are 13 or 14, so the teen range fits. That is why our two-way tables sort sleep into 8 or more hours and fewer than 8 hours.
Source: Centers for Disease Control and Prevention, About Sleep (cdc.gov)