Inference for means: mixed t procedures worksheet
8 problems on inference for means: mixed t procedures, ready to print for class or homework. The answer key prints on its own page, so hand out everything before it.
Name: ______________________________ Date: ______________ Period: ______
Problem 1
Name the procedure each study calls for, and give the reason. Do not compute anything.
(a) A random sample of 40 used cars is tested, and the goal is to estimate the mean carbon dioxide output per kilometer with 95% confidence. The population standard deviation is unknown. (b) Twenty-five volunteers each taste two formulas of the same soda and rate both on a 100-point scale. (c) An independent random sample of 30 seniors and an independent random sample of 30 first-year students report nightly sleep, and the question is whether mean sleep differs. (d) A random sample of 18 batteries is tested against the manufacturer's claim that mean life is 500 hours.
Problem 2
A 3D printer's manufacturer advertises a mean print time of 45 minutes for one benchmark model. A maker space times a random sample of 20 prints of that model on its own machine and finds minutes with minutes; a dotplot of the 20 times is roughly symmetric with no outliers. At , is there convincing evidence that this machine's true mean print time differs from the advertised 45 minutes?
Problem 3
A biologist measures forewing length, in millimeters, on an independent random sample of 15 butterflies from Meadow Site and an independent random sample of 15 from Woodland Site. Meadow gives with ; Woodland gives with . A student says that because both samples have 15 observations, the data can be paired. (a) Explain why that is wrong and name the correct procedure. (b) Carry it out at using the conservative degrees of freedom. (c) State what the study design does and does not let the biologist claim.
Problem 4
An environmental lab checks two chlorine test kits by running both kits on each of 12 randomly selected water samples. For the differences (Kit 1 reading minus Kit 2 reading), in parts per million, the lab reports and . (a) Name the procedure and say what feature of the design fixes it. (b) Construct a 95% confidence interval for the true mean difference. (c) Interpret the interval and say what it implies about whether the kits agree.
Problem 5
A poultry researcher compares two feed formulas. An independent random sample of 14 hens on Feed A lays a mean of 232 eggs per year with ; an independent random sample of 16 hens on Feed B lays a mean of 221 eggs with . Construct a 90% confidence interval for the difference in true mean annual egg production using the conservative degrees of freedom, and interpret it.
Problem 6
A public health guideline says a single frozen dinner should average no more than 600 calories. A researcher measures a random sample of 22 frozen dinners from one manufacturer and finds calories with calories, with a boxplot showing no outliers and little skew. A student proposes a two-sample test, treating the 22 dinners as one sample and the 600-calorie guideline as the other. (a) Explain what is wrong with that and name the correct procedure. (b) Carry it out at . (c) Interpret the p-value in context.
Problem 7
Two weather stations report weekly rainfall, in millimeters. From an independent random sample of 12 weeks at Station A and an independent random sample of 15 weeks at Station B, an analyst reports a 95% confidence interval for of , built with the conservative degrees of freedom. (a) Name the procedure and give the degrees of freedom and used. (b) Recover the observed difference in sample means, the margin of error, and the standard error. (c) State what a two-sided test of at would conclude, without computing a p-value. (d) A colleague writes that the study proves the two stations get the same mean weekly rainfall. Correct that.
Problem 8
A teacher wants to know whether reading medium affects comprehension. Two designs are proposed.
Design A. Twenty students are randomly assigned, 10 to read a passage on paper and 10 to read a different but comparable passage on screen. Design B. Ten students each read two comparable passages, one on paper and one on screen, with the order and the passage assignment randomized for each student.
(a) Name the procedure each design calls for. (b) Design B is run, and the differences (paper score minus screen score) are . Test at whether reading medium affects mean comprehension. (c) Explain why Design B is more likely than Design A to detect a real difference of this size. (d) State the scope of inference.
From statslearn.org/worksheets/mean-inference-mixed-worksheet. Free to print and copy for classroom use.
Inference for means: mixed t procedures: answer key
(a) One-sample interval, . (b) Paired procedure on 25 differences, . (c) Two-sample test, by hand. (d) One-sample test against , .
One-sample test. , , , two-sided p-value (table bracket to ). Reject at ; convincing evidence the true mean print time differs from 45 minutes, though the result would not reject at .
(a) Different butterflies with no link between individuals, so the data are two independent samples; equal sizes do not create pairing. (b) mm, , conservative , two-sided p-value (table bracket to ). Reject ; convincing evidence the true mean forewing lengths differ. (c) Random selection generalizes to each site, but with no random assignment the difference cannot be attributed to habitat.
(a) Paired interval, because both kits measure the same 12 water samples, so pairs and . (b) ppm, , , interval ppm. (c) 95% confident it captures the true mean difference; it excludes 0, so Kit 1 reads higher on average, though the difference may be too small to matter in practice.
Two-sample interval. eggs, , conservative , , . The 90% interval is eggs per year. It contains 0, so there is no convincing evidence of a difference, and the interval is too wide to settle the question either way.
(a) A published guideline is a number, not a sample, so it carries no sampling variability; use a one-sample test with . (b) , , , one-sided p-value (table bracket to ). Reject ; convincing evidence the true mean exceeds 600 calories. (c) If the true mean were 600, only about 2.07% of samples of 22 would average 638 calories or more.
(a) Two-sample interval, conservative , . (b) mm, mm, mm. (c) The interval contains 0, so the matching two-sided test at fails to reject . (d) Failing to detect a difference is not proof of no difference; a gap as large as 4 mm is still a plausible value.
(a) Design A: two-sample test, conservative . Design B: paired test on 10 differences, . (b) , , , , , two-sided p-value (below from the table); reject , and the 95% interval agrees. (c) Pairing removes between-student ability differences from the standard error, so a 2-point effect stands out. (d) Causal for these students because of random assignment, but not generalizable, because they were not randomly selected.
Every step is worked at statslearn.org/practice/mean-inference-mixed-practice. Free to print and copy for classroom use.
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
- One-sample t interval worksheet with answer key
- Difference of two sample means worksheet with answer key
- Justifying claims from a mean CI: worksheet with answer key
- Paired t test worksheet with answer key
- Sampling distribution of x-bar: 8 worksheet with answer key
- t-test worksheet with answer key
- Two-sample t test worksheet with answer key