write the steps for making an inference from a random sample, and report an estimate with a likely range instead of a single exact number.
Phase 3: Make Sense
Do Now. Sample data: our class took 8 random samples of 40 students from a school of 480. The percent who prefer a Saturday carnival was 30%, 35%, 32.5%, 40%, 35%, 37.5%, 35% and 32.5%. What single number would you report? What range?
Which report is the most honest?
The reading opens over the lesson; close it to come back to exactly where you were.
Our method has a name. Using a sample to draw a conclusion about a population is called statistical inference. We have built it piece by piece; today we put the steps in order.
The five steps. (1) Name the population and the question. (2) Make a complete list and choose a random sample. (3) Find the sample result, such as a percent or a mean. (4) Use proportional reasoning to estimate for the population. (5) Report how sure you are, using the spread of repeated samples as a likely range.
Worked example 1: the Saturday carnival. Sample data from the Do Now. Step 1: the 8 sample percents add to 277.5, so the mean is 277.5 ÷ 8 = 34.6875%, about 35%. Step 2: the lowest is 30% and the highest is 40%. Step 3: scale to 480 students: 35% of 480 = 168; 30% of 480 = 144; 40% of 480 = 192. Step 4: report: “About 168 students, likely between 144 and 192, would come on Saturday.”
Worked example 2: reading a Census range. Hana compares the two intervals in the reading. Step 1: 90% width: 37,315,094 − 35,534,124 = 1,780,970. Step 2: 95% width: 37,485,612 − 35,363,606 = 2,122,006. Step 3: difference: 2,122,006 − 1,780,970 = 341,036. Step 4: the 95% interval is wider by 341,036 people, because being more confident needs a wider net.
Our range is simpler. The Census Bureau computes its intervals with formulas you will meet in high school. In grade 7 we use the lowest and highest results of our repeated samples. That is a fair, honest way to show how much our estimate could move.
Claim, evidence, reasoning. In our report, each estimate is a claim. The evidence is the sample data. The reasoning explains why the sample was random and how we scaled it. A reader should be able to check every step.
Use the five steps. Show each step for full credit.
Apply the method.
Sample data: 6 random samples of 30 students from a school of 600 found 9, 12, 10, 11, 8 and 10 students who want a dance. Find the mean percent, then report an estimate and a likely range for the school.
Sample data: a report says “about 168 students, likely between 144 and 192.” How wide is that range?
Explain why a range from 0% to 100% is always correct but never useful.
We named our method. Next we name the method for comparing two groups: is the gap big compared with the spread?
Exit ticket: list the five steps of inference from memory.
Pacing: Do Now 5 min, reading 7, mini lesson 10, guided practice 13, independent practice and exit ticket 5.
Accuracy: the definition of a confidence interval, the 1995 poverty example from the March 1996 Current Population Survey, the 90% interval 35,534,124 to 37,315,094, the 95% interval 35,363,606 to 37,485,612, the repeated-sample meaning, the Bureau’s routine use of 90% and the caution statement are from the Census page opened 2026-11-20. Our min-to-max range of class samples is a grade 7 stand-in, not a confidence interval; say so if students ask. Checked: the eight percents sum to 277.5, mean 34.6875; 30%, 35%, 40% of 480 = 144, 168, 192; widths 1,780,970 and 2,122,006, difference 341,036; 192 − 144 = 48.
Likely misconceptions: reporting the estimate as exact; thinking a 95% interval is narrower because 95 is bigger than 90; skipping the randomness check.
Supports: a five-step card with blanks. Stretch: use the Census 90% interval to find its midpoint. ((35,534,124 + 37,315,094) ÷ 2 = 36,424,609.)
Answer guide: dance data: counts 9, 12, 10, 11, 8, 10 of 30 are 30%, 40%, 33 1/3%, 36 2/3%, 26 2/3%, 33 1/3%; total count 60 of 180 = 33 1/3%; about 200 of 600 students, likely between 160 (26 2/3%) and 240 (40%). The CER paragraph should name the random method, the mean of 35% and the range 144 to 192.
U.S. Census Bureau, A Basic Explanation of Confidence Intervals · U.S. government work, public domain
The U.S. Census Bureau often reports an estimate together with a confidence interval. A confidence interval is a range of values that shows how uncertain an estimate is. It is written with two endpoints.
The page gives a real example. For the number of people of all ages in poverty in the United States in 1995, based on a survey taken in March 1996, the 90% confidence interval was 35,534,124 to 37,315,094 people.
What does the 90% mean? If the Bureau repeated the same steps again and again, drawing a new sample each time, about 90% of the intervals made this way would contain the true value. The Bureau usually uses 90% intervals.
To be more confident, the interval has to be wider. For the same estimate, the 95% confidence interval was 35,363,606 to 37,485,612 people.
The page says confidence intervals are a reminder that estimates have limits. The wider the interval for an estimate, the more careful people need to be when they use it.
Source: U.S. Census Bureau, A Basic Explanation of Confidence Intervals (census.gov)