Checking conditions for inference worksheet with answer key

8 problems on checking conditions for inference, ready to print for class or homework. The answer key prints on its own page, so hand out everything before it.

Problem 1

A regional transit agency states that 35% of its riders use a monthly pass. The agency serves about 24,000 riders. A researcher takes a random sample of 240 riders and finds 96 with a monthly pass. Check every condition for a one-sample z-test of the null hypothesis H0:p=0.35H_0: p = 0.35 against the alternative hypothesis Ha:p≠0.35H_a: p \ne 0.35, showing the numbers.

Problem 2

A seed library has about 900 members. Staff take a random sample of 60 members and find that 8 of them returned seeds at the end of the season. (a) Are the conditions met for a one-sample z-test of H0:p=0.20H_0: p = 0.20 against Ha:p<0.20H_a: p < 0.20? (b) Are the conditions met for a one-sample z-interval for pp? (c) Explain why the two answers differ.

Problem 3

A hospital employs 3,200 people and wants a 95% confidence interval for the mean one-way commute time of its employees. A random sample of 46 employees gives xˉ=27.4\bar{x} = 27.4 minutes (read 'x-bar', the sample mean) and s=9.6s = 9.6 minutes (the sample standard deviation). Nothing is known about the shape of the population of commute times. Check the conditions for a one-sample t-interval, then find the standard error.

Problem 4

A bookshop owner wants a confidence interval for the mean time a customer spends in the shop. She randomly selects 14 of the roughly 300 customers from one week and records their times. (a) The dotplot of the 14 times is roughly symmetric with no outliers. Are the conditions for a one-sample t-interval met? (b) Suppose instead the dotplot is strongly right-skewed with one customer at 95 minutes, far above the rest. Are the conditions met? (c) What would change your answer in (b)?

Problem 5

A city tests three recycling-bin labels. Sixty volunteers are randomly assigned to each label, 180 in all, and each volunteer sorts one bag of mixed waste. The number who sort correctly is 39 for Label A, 45 for Label B, and 30 for Label C.

LabelSorted correctlyNot sorted correctlyRow total
A392160
B451560
C303060
Total11466180

Check the conditions for a χ2\chi^2 test for homogeneity, showing every expected count.

Problem 6

A city runs a randomized experiment at its parking meters. Of 180 drivers randomly assigned a paper receipt prompt, 117 pay for the full time; of 180 drivers randomly assigned a text prompt, 135 pay for the full time. Let p1p_1 and p2p_2 be the true full-payment proportions for the paper and text prompts. (a) Check the conditions for a two-sample z-test of H0:p1=p2H_0: p_1 = p_2. (b) Check the conditions for a two-sample z-interval for p1−p2p_1 - p_2. (c) State which counts each procedure uses and why.

Problem 7

A podcast host reads a link on air and asks listeners to visit a page and say whether they skip the ads. 1,500 listeners respond and 1,080 say they skip. The host wants to test H0:p=0.65H_0: p = 0.65 against Ha:p>0.65H_a: p > 0.65 for all of the show's listeners, of whom there are several million. (a) Check each condition. (b) Is the test justified? (c) The host offers to leave the page open until 15,000 people respond. Does that fix the problem?

Problem 8

A county fair drew about 18,000 attendees. Organizers survey a random sample of 150 of them. Of those surveyed, 27 arrived on the free shuttle and 123 did not. The age breakdown of the 150 is: under 18, 30 people; 18 to 59, 96 people; 60 and over, 24 people. (a) Are the conditions met for a one-sample z-interval for the proportion of attendees who arrived on the shuttle? (b) Organizers also want a χ2\chi^2 test for independence between age group and whether the attendee took the shuttle. Are those conditions met? (c) Explain how the same 150 responses can justify one procedure and not the other, and say what would fix (b).

Checking conditions for inference: answer key

  1. Random: stated random sample of riders. 10%: 240≤0.10(24000)=2400240 \le 0.10(24000) = 2400. Large Counts: np0=84np_0 = 84 and n(1−p0)=156n(1 - p_0) = 156, both at least 10. All three met, so the one-sample z-test is justified.

  2. (a) Yes: np0=12np_0 = 12 and n(1−p0)=48n(1 - p_0) = 48, both at least 10, so the test is justified. (b) No: the interval uses observed counts and np^=8n\hat{p} = 8 is below 10. (c) A test checks the counts at p0p_0, while an interval has no p0p_0 and must check them at p^\hat{p}.

  3. All three met: random sample; 46≤0.10(3200)=32046 \le 0.10(3200) = 320; and n=46≥30n = 46 \ge 30, so no plot is needed. The one-sample t-interval is justified with df=45df = 45 and SE=9.646≈1.415SE = \frac{9.6}{\sqrt{46}} \approx 1.415 minutes.

  4. (a) Met: random, 14≤3014 \le 30, and with n<30n < 30 a roughly symmetric plot with no outliers satisfies the sample data condition. (b) Not met: strong right skew and an outlier at n=14n = 14, so the t-interval is not justified. (c) A sample of at least 30, or evidence that the population is approximately normal.

  5. Random: randomized experiment. 10%: not needed for a randomized experiment. Expected counts are 38 in each 'sorted correctly' cell and 22 in each 'not sorted correctly' cell, all greater than 5. The chi-square test for homogeneity is justified.

  6. (a) Justified: randomized experiment, and the pooled p^c=0.70\hat{p}_c = 0.70 gives expected counts 180(0.70)=126180(0.70) = 126 and 180(0.30)=54180(0.30) = 54 in each group, all at least 10. (b) Justified: the observed counts 117, 63, 135, and 45 are all at least 10. (c) The test uses pooled counts, the interval uses each group's own observed counts.

  7. (a) 10% and Large Counts pass easily (np0=975np_0 = 975 and n(1−p0)=525n(1 - p_0) = 525), but Random fails: this is a voluntary response sample. (b) Not justified. (c) No, a larger voluntary response sample shrinks the standard error without touching the bias.

  8. (a) Met: random sample, 150≤1800150 \le 1800, and observed counts 27 and 123 are both at least 10, so the z-interval is justified. (b) Not met: the expected count for '60 and over, shuttle' is 24(27)150=4.32\frac{24(27)}{150} = 4.32, which is not greater than 5, so the chi-square test is not justified. (c) Conditions are checked per procedure; a larger sample, or more attendees sampled from the 60 and over group, would fix it.

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