Justifying claims: confidence interval worksheet

8 problems on justifying claims: confidence interval, ready to print for class or homework. The answer key prints on its own page, so hand out everything before it.

Problem 1

A town with about 40,000 registered voters draws a random sample of 900 of them and finds 486 who support a bond measure. The 95% confidence interval for pp, the true proportion of all registered voters in the town who support the measure, is (0.507, 0.573)(0.507,\ 0.573).

(a) Does the interval support the claim that more than half of all registered voters support the measure? (b) A council member says the interval shows support is 54%. Is that claim supported? (c) A second council member writes, "There is a 95% chance the true proportion is between 0.507 and 0.573." Name the error and fix the sentence.

Problem 2

A district with 6,200 students takes a random sample of 400 of them and finds 92 who walk to school.

(a) Check the conditions and construct the 95% confidence interval for the proportion of all students in the district who walk to school. (b) The district's transportation plan assumes at least 30% of students walk. Does the interval give convincing evidence against that assumption? (c) A parent group claims the figure is about 25%. Does the interval give convincing evidence against that claim? (d) Write the sentence that justifies your answer to (b).

Problem 3

A county with 21,000 households draws a random sample of 500 of them and finds 220 that compost food waste. A county report claims 48% of households compost; a competing report claims half do.

(a) Build the 90% confidence interval and use it to judge both claims. (b) Build the 99% confidence interval from the same sample and judge both claims again. (c) One claim changed verdict. Explain what that shows, and what it does not show.

Problem 4

A clinic randomly assigns 250 seasonal allergy patients to two inhalers, 125 to a new formulation and 125 to the standard one. After four weeks, 95 patients on the new formulation and 88 on the standard report no symptoms. With the difference defined as p1−p2p_1 - p_2, new minus standard, the 95% confidence interval is (−0.054, 0.166)(-0.054,\ 0.166).

(a) What does this interval let the clinic claim? (b) A summary of the study says "we accept the null hypothesis that the two inhalers work equally well." Give the two errors in that sentence. (c) A second summary says the study shows the new inhaler is no better. Is that supported?

Problem 5

Two campuses of one university are compared on the proportion of students who use the free tutoring center. An independent random sample of 400 students from the North campus (enrollment 9,000) contains 168 users. An independent random sample of 450 students from the South campus (enrollment 12,000) contains 252 users. With the difference defined as pN−pSp_N - p_S, the 95% confidence interval is (−0.207, −0.073)(-0.207,\ -0.073).

(a) What claim does this interval justify? (b) A student says the upper endpoint −0.073-0.073 is so close to 0 that a difference of 0 is still plausible. Is that right? (c) What would the interval and the conclusion be if the difference had been defined as pS−pNp_S - p_N?

Problem 6

Independent random samples of 1,000 adults are taken in each of two large cities and asked whether they hold a public library card. City A returns 570 card holders and City B returns 520. The 95% confidence interval for pA−pBp_A - p_B is (0.006, 0.094)(0.006,\ 0.094). Say which of the five statements are correct, and for each wrong one name the specific error.

  1. There is a 95% chance the true difference is between 0.006 and 0.094.
  2. Because 0 is outside the interval, a two-sided test of H0:pA=pBH_0: p_A = p_B at α=0.05\alpha = 0.05 would reject H0H_0.
  3. The interval shows the true difference is 0.05.
  4. Because 0 is outside the interval, the probability that H0H_0 is true is less than 0.05.
  5. We are 95% confident that the interval from 0.006 to 0.094 contains the true difference in the proportions of adults holding a library card, City A minus City B, so there is convincing evidence that City A's rate is higher.

Problem 7

A company with 20,000 employees runs a voluntary wellness program that about 6,000 employees have joined. An independent random sample of 300 enrolled employees contains 195 who met the annual step goal. An independent random sample of 400 employees who are not enrolled contains 200 who met it.

(a) Construct the 99% confidence interval for p1−p2p_1 - p_2, enrolled minus not enrolled. (b) Write the sentence that justifies a claim from it. (c) The company newsletter says the program raises the goal-meeting rate by 15 percentage points. Is that supported? (d) Does the interval support the claim that the program causes higher rates?

Problem 8

An online store randomly assigns 10,000 visitors to two checkout pages, 5,000 to each. Checkout is completed by 415 visitors on the new page and 350 on the old page. With the difference defined as p1−p2p_1 - p_2, new minus old, the 95% confidence interval is (0.0026, 0.0234)(0.0026,\ 0.0234).

(a) Does the interval support the claim that the new page increases the completion rate? (b) A product manager wants to claim the new page raises the rate by at least 5 percentage points. Does the interval support that? (c) The manager replies that the result is statistically significant, so the improvement must be large. Answer that.

Justifying claims: confidence interval: answer key

  1. (a) Yes. The whole interval lies above 0.50, so there is convincing evidence that more than half of the town's registered voters support the measure. (b) No. 0.54 is the sample proportion and only one of many plausible values; an interval never singles out one number as the truth. (c) A computed interval carries no probability. Write "we are 95% confident that the interval from 0.507 to 0.573 contains the true proportion of all registered voters in this town who support the bond measure."

  2. (a) About (0.189, 0.271)(0.189,\ 0.271), with all three conditions met. (b) Yes. 0.30 lies above the whole interval. (c) No. 0.25 is inside the interval, so it stays plausible. (d) "We are 95% confident that the interval from 0.189 to 0.271 contains the true proportion of students in this district who walk to school. Because 0.30 is not in that interval, there is convincing evidence that fewer than 30% of the district's students walk to school."

  3. (a) 90%: about (0.403, 0.477)(0.403,\ 0.477). Both 0.48 and 0.50 lie above it, so there is convincing evidence against both claims. (b) 99%: about (0.383, 0.497)(0.383,\ 0.497). Now 0.48 is inside and stays plausible, while 0.50 is still above the interval. (c) A higher confidence level widens the interval and rules out fewer values, so a justification means nothing without its level attached; the data did not change and neither interval is wrong.

  4. (a) Only that 0 is a plausible difference, so there is not convincing evidence the two inhalers differ; the interval spans negative and positive values, so it names no direction either. (b) You never accept a null hypothesis, and "work equally well" singles out 0 from a whole range of plausible differences; a real difference this study missed would be a Type II error. (c) No. Values up to 0.166 stay plausible, so an advantage for the new inhaler is not ruled out.

  5. (a) That the South campus has the higher true tutoring rate, by plausibly 7.3 to 20.7 percentage points, since both endpoints are negative and 0 is outside. (b) No. 0 lies above −0.073-0.073, so it is outside the interval; endpoints are compared as signed numbers. (c) (0.073, 0.207)(0.073,\ 0.207), and the conclusion is unchanged: the South campus rate is higher.

  6. Statements 2 and 5 are correct. Statement 1 attaches a probability to a finished interval, statement 3 promotes the point estimate 0.05 to the truth, and statement 4 treats a significance level as the probability that H0H_0 is true, when a p-value is computed assuming H0H_0 holds and α\alpha is the long-run error rate of the procedure.

  7. (a) About (0.054, 0.246)(0.054,\ 0.246), with SE≈0.037193SE \approx 0.037193 and ME≈0.0958ME \approx 0.0958. (b) We are 99% confident that the interval from 0.054 to 0.246 contains the true difference in goal-meeting proportions, enrolled minus not enrolled; because 0 is outside and both endpoints are positive, there is convincing evidence enrolled employees meet the goal more often. (c) No. 0.15 is the point estimate; the interval supports everything from 0.054 to 0.246. (d) No. Employees self-selected, so the data are observational and cannot establish causation.

  8. (a) Yes. Both endpoints are positive, so 0 is not plausible, and random assignment lets the store attribute the increase to the page for the visitors in this experiment. (b) No. 0.05 lies above the whole interval, so there is convincing evidence against a gain of at least 5 percentage points; plausible gains run from about 0.26 to 2.34 percentage points. (c) Significance says the difference is not 0, not that it is large. The interval bounds the size, and here it caps the gain below 2.4 percentage points.

The key gives final answers only. The full worked solution for each problem is on the matching practice set, and every other sheet is on the worksheets page.

More worksheets from this unit