AP Stats Exam - Multiple Choice & Free Response

Apr 2, 2026

All files

summarize_document

Here's a summary of the provided content, organized by question:

Summary of AP Statistics Practice Exam Questions

This document contains a series of multiple-choice and free-response questions from an AP Statistics practice exam, along with answer keys and scoring guidelines. The questions cover a range of statistical concepts, including:

Descriptive Statistics (histograms, boxplots, stemplots, scatterplots)
Probability and Simulation
Sampling and Experimentation
Inferential Statistics (confidence intervals, hypothesis testing, p-values, z-tests, t-tests, chi-square tests)

Multiple-Choice Questions (Items 13-40)

13. Bear Ages and Sex:

Topic: Data analysis using a back-to-back stemplot.
Task: Determine the true statement comparing the median age and range of ages for male and female bears based on the provided stemplot.

14. Produce Box Weights:

Topic: Normal distribution and percentiles.
Task: Identify the expression representing the weight at the 75th percentile for a distribution with a given mean and standard deviation. This involves understanding that the 75th percentile corresponds to a specific z-score.

15. Confidence Interval Interpretation:

Topic: Understanding confidence levels in statistical inference.
Task: Choose the best interpretation of a 95 percent confidence level for a proportion, given multiple confidence intervals constructed from samples.

16. Difference in Proportions Standard Error:

Topic: Calculating the standard error for the difference between two proportions.
Task: Identify the correct formula for the standard error of the difference in proportions of adults sleeping at least seven hours, based on sample data for exercisers and non-exercisers.

17. Hypothesis Testing and p-value:

Topic: Decision making in hypothesis testing using a p-value.
Task: Determine if there is sufficient evidence to conclude that the proportion of Americans owning American-made cars has increased, given a sample proportion, a historical proportion, and a p-value, at a significance level of $\alpha = 0.05$.

18. Random Assignment in Experiments:

Topic: Principles of experimental design.
Task: Identify the most important reason for randomly assigning colors to roles (helper/hinderer) in a study investigating babies' sense of right and wrong.

19. Sodium Content of Potato Chips:

Topic: Comparing distributions using boxplots.
Task: Make a correct comparison between the sodium content of "healthy" and "regular" potato chips based on provided boxplots, considering measures like IQR, range, and median.

20. Defective Parts from Machines:

Topic: Analyzing relationships using a scatterplot.
Task: Describe the relationship between the number of defective parts produced by Machine A and Machine B over 19 days, based on a scatterplot.

22. Sum of Dice Rolls:

Topic: Probability distributions and random variables.
Task: Determine the probability distribution for the sum of points earned from two independent tosses of a four-sided die.

23. Confidence Interval for Population Mean:

Topic: Choosing the appropriate confidence interval procedure.
Task: Identify the appropriate method for constructing a one-sample confidence interval for the population mean surface area of maple leaves, given that the population distribution is normal and the sample standard deviation is known.

24. Comparing Teacher Salaries:

Topic: Hypothesis testing for the difference between two means.
Task: Determine if there is convincing statistical evidence that teacher salaries in region A are greater than in region B, based on sample data and significance levels.

25. Experimental Design:

Topic: Identifying experimental designs.
Task: Describe the florist's method of assigning roses to water treatments (with or without powder) as either a census, experiment (completely randomized, blocked, matched-pairs), or observational study.

26. Sample Size for Margin of Error:

Topic: Calculating sample size for a confidence interval.
Task: Find the smallest sample size required to achieve a specific margin of error (at most 0.02) at a 95 percent confidence level for estimating a proportion.

27. Probability of Final Winnings:

Topic: Calculating probabilities involving multiple stages and random variables.
Task: Determine the probability that a contestant's final winnings are exactly $2,000, considering the selection of a base amount and a multiplier chip.

28. Regression Confidence Interval Interpretation:

Topic: Interpreting confidence intervals for the slope of a regression line.
Task: Choose the correct interpretation of a 95 percent confidence interval for the slope of the least-squares regression line relating maximum wind speed to minimum pressure of hurricanes.

29. Study Design Flaws:

Topic: Identifying limitations in study design.
Task: Explain why a study on the effectiveness of a new eye-drop solution for contact lens wearers does not support the conclusion that it is more effective than no solution.

30. Confidence Interval for Proportion:

Topic: Constructing a confidence interval for a population proportion.
Task: Identify the correct formula for a 99 percent confidence interval for the proportion of all employees who own an electric car, based on a sample.

31. p-value from t-statistic:

Topic: Hypothesis testing and p-values.
Task: Estimate the p-value for a one-sided t-test given the sample size and the test statistic.

32. Comparing Distributions from a Graph:

Topic: Interpreting bar graphs showing distributions across categories.
Task: Identify a statement that must be true based on a graph showing the distribution of volunteer activities chosen by students in grades 6, 7, and 8.

33. Interpretation of "Strong" Association:

Topic: Understanding correlation and association in scatterplots.
Task: Interpret the meaning of "strong" in the context of a strong, positive linear association between baggage items and delivery time.

34. Choosing a Statistical Test:

Topic: Selecting appropriate statistical tests for association.
Task: Identify the most appropriate statistical procedure for investigating an association between gender and the number of social media friends.

35. Probability with Multiple Distributions:

Topic: Calculating probabilities involving a mixture of distributions.
Task: Estimate the probability that a commute time will be greater than 25 minutes, given different normal distributions for commute times under good and not-good weather conditions, and the probability of good weather.

36. Outlier Definition and Distribution Shape:

Topic: Identifying outliers and comparing mean and median.
Task: Conclude whether the median is greater or less than the mean, and determine the number of outliers based on a given definition and summary statistics/histogram of the number of siblings.

37. Hypothesis Testing for Proportions:

Topic: Drawing conclusions from a two-sample z-test for proportions.
Task: Make a correct conclusion about the difference in proportions of overweight dogs in Florida and Colorado, given a z-test statistic and a significance level.

38. Stratified Sampling Design:

Topic: Understanding sampling methods.
Task: Identify the method that achieves a stratified sampling design based on voting districts.

39. Simulation of Sampling Distribution:

Topic: Understanding sampling distributions.
Task: Describe the distribution being simulated by rolling 10 dice and recording the average of the numbers, repeated 20 times.

40. Independence and Mutual Exclusivity:

Topic: Concepts of probability, independence, and mutual exclusivity.
Task: Determine whether two events (being awarded electrical wiring jobs) are mutually exclusive and/or independent, given their individual and joint probabilities.

Free-Response Questions (Items 1-6)

Question 1: Natural Gas Usage and Temperature

Topic: Data visualization and interpretation of relationships.
Parts:
- (a) Describe the trend of natural gas usage over the year.
- (b) Construct an appropriate graph (scatterplot) to show the relationship between therms used and average monthly temperature.
- (c) Describe the relationship revealed by the scatterplot that is not evident from the bar chart.

Question 2: Chocolate Consumption and Stroke

Topic: Hypothesis testing and observational studies.
Parts:
- (a) Perform a hypothesis test to determine if men who eat the most chocolate have a lower probability of stroke than those who eat the least.
- (b) Evaluate the appropriateness of a newspaper's conclusion that eating more chocolate reduces stroke probability, justifying the answer based on the study design.

Question 3: Employee Benefits Survey

Topic: Sampling methods (stratified vs. cluster sampling).
Parts:
- (a) Identify which sampling method is stratified and which is cluster, and identify the strata and clusters.
- (b) Provide a statistical advantage and disadvantage of cluster sampling (method 1).
- (c) Explain a statistical advantage of stratified sampling (method 2) compared to simple random sampling.

Question 4: Diversity Index

Topic: Probability calculations and maximizing functions.
Parts:
- (a) Calculate the diversity index D when all organisms are the same species.
- (b) Calculate D for a population with known proportions of two species.
- (c) Determine the maximum possible value of D for a population with two species, justifying the answer.

Question 5: City Park Proposal Support

Topic: Binomial distribution, normal approximation, and inference.
Parts:
- (a) Calculate the mean and standard deviation for the number of residents supporting a proposal, assuming a binomial distribution.
- (b) Use a normal approximation to find the interval (k1, k2) representing the middle 0.997 of the distribution.
- (c) Use the survey results and the interval from (b) to determine if the proposal should be funded, justifying the answer.

Question 6: Customer Wait Times

Topic: Confidence intervals, data transformation, and interpretation.
Parts:
- (a) Explain why a one-sample t-interval might not be appropriate for raw wait time data based on a histogram.
- (b) Construct and interpret a 95 percent confidence interval for the mean of log-transformed wait times, given that conditions are met.
- (c) Convert the endpoints of the confidence interval from log minutes back to minutes.
- (d) Compare the parameter $10^\mu$ to the median and mean of the population distribution of wait times.
- (e) Interpret the confidence interval constructed in part (c) in the context of wait times.

Answer Key and Scoring Guidelines

The document also includes:

A multiple-choice answer key.
Scoring guidelines for the free-response questions, detailing the criteria for "Essentially Correct (E)," "Partially Correct (P)," and "Incorrect (I)" for each part.
A scoring worksheet and conversion chart for calculating composite scores.
Question descriptors and performance data indicating the statistical concepts assessed and student performance on each question.

Summary of AP Statistics Practice Exam Content

This document contains materials related to the AP Statistics Exam, including free-response questions, scoring guidelines, and multiple-choice answer keys. The content covers various statistical concepts and their applications.

Section II: Free-Response Questions

This section includes detailed questions and scoring guidelines for the free-response portion of the exam.

Question 1: Comparing Distributions (Boxplots)

Scenario: Comparing the distribution of yearly salaries for entry-level accountants at two corporations (A and B) five years after hiring. Data is presented using boxplots.
Part (a): Compare the distributions of salaries for Corporation A and Corporation B, focusing on center, spread, and outliers.
- Key Findings:
  - Median salaries are similar for both corporations.
  - Corporation A shows greater variability (larger range and IQR) than Corporation B.
  - Corporation A has outliers representing very high salaries, while Corporation B has no outliers.
Part (b): Provide reasons for choosing a job offer from either Corporation A or Corporation B, based on the boxplots.
- Reason to choose Corporation A: Potential for higher salary due to higher outliers and greater spread.
- Reason to choose Corporation B: Higher minimum salary and less variability, suggesting more consistent raises and a higher guaranteed minimum.

Question 2: Confidence Intervals for Proportions

Scenario: A restaurant owner is concerned about a discount program's probability of generating discounts. The program is intended to give a discount with a probability of 0.2. A sample found 15% received discounts, leading to a confidence interval of 0.15 ± 0.06.
Part (a): Interpret the confidence interval (0.09, 0.21).
- (i) Evidence of program not working as intended: No. The interval includes 0.20, so it's plausible the program is working correctly.
- (ii) Evidence of program generating discount with probability 0.2: No. The interval includes values other than 0.20, meaning other probabilities are also plausible.
Part (b): Calculate the new margin of error if the sample size is quadrupled, keeping the sample proportion at 0.15.
- Method: The margin of error is inversely proportional to the square root of the sample size. Quadrupling the sample size reduces the margin of error by a factor of $\sqrt{4} = 2$.
- New Margin of Error: (0.06) / 2 = 0.03.
Part (c): Conclude about the program's intended function based on the new interval (0.15 ± 0.03 = (0.12, 0.18)).
- Conclusion: Yes, there is convincing evidence the program is not working as intended because the new interval does not contain 0.20.

Question 3: Discrete Probability Distributions

Scenario: A shopping mall has three ATMs. Let X be the number of ATMs working when the mall opens. A probability distribution for X is provided.
Part (a): Calculate the probability that at least one ATM is working ($P(X \ge 1)$).
- Calculation: $P(X \ge 1) = P(X=1) + P(X=2) + P(X=3) = 0.21 + 0.40 + 0.24 = 0.85$. Alternatively, $1 - P(X=0) = 1 - 0.15 = 0.85$.
Part (b): Calculate the expected value of the number of working ATMs ($E(X)$).
- Calculation: $E(X) = (0 \times 0.15) + (1 \times 0.21) + (2 \times 0.40) + (3 \times 0.24) = 0 + 0.21 + 0.80 + 0.72 = 1.73$.
Part (c): Calculate the conditional probability that all three ATMs are working given at least one is working ($P(X=3 | X \ge 1)$).
- Calculation: $P(X=3 | X \ge 1) = P(X=3 \text{ and } X \ge 1) / P(X \ge 1) = P(X=3) / P(X \ge 1) = 0.24 / 0.85 \approx 0.282$.
Part (d): Compare the expected value of working ATMs given at least one is working to the original expected value.
- Conclusion: The conditional expected value will be greater than the original expected value.
- Explanation: Eliminating the case where zero ATMs are working increases the probabilities of the remaining cases (1, 2, or 3 working ATMs), shifting the distribution's balance point higher.

Question 4: Hypothesis Test for Proportions (Two-Sample Z-Test)

Scenario: Investigating if low-dose aspirin reduces the chance of developing colon cancer. 1000 volunteers were randomly assigned to an aspirin group (500) or a placebo group (500). Results: 15 aspirin users developed cancer, 26 placebo users developed cancer. Significance level $\alpha = 0.05$.
Hypotheses:
- $H_0: p_{aspirin} = p_{placebo}$ (Aspirin does not reduce the chance of colon cancer).
- $H_a: p_{aspirin} < p_{placebo}$ (Aspirin reduces the chance of colon cancer).
Test Procedure: Two-sample z-test for proportions.
Conditions: Random assignment (satisfied), large sample counts (all counts > 5, satisfied).
Calculations:
- Sample proportions: $\hat{p}{aspirin} = 15/500 = 0.030$, $\hat{p}{placebo} = 26/500 = 0.052$.
- Combined proportion: $\hat{p}_{combined} = (15+26)/(500+500) = 41/1000 = 0.041$.
- Test statistic: $z = (\hat{p}{aspirin} - \hat{p}{placebo}) / \sqrt{\hat{p}{combined}(1-\hat{p}{combined})(1/n_1 + 1/n_2)} = (0.030 - 0.052) / \sqrt{0.041(1-0.041)(1/500 + 1/500)} \approx -1.75$.
- p-value: $P(Z \le -1.75) \approx 0.0401$.
Conclusion: Since the p-value (0.0401) is less than $\alpha = 0.05$, reject $H_0$. There is convincing statistical evidence that taking low-dose aspirin daily reduces the chance of developing colon cancer.

Question 5: Linear Regression and Scatterplots

Scenario: Measuring heights and arm spans of 12 seniors. A scatterplot is provided.
Part (a): Describe the relationship between arm span and height based on the scatterplot.
- Description: Moderately strong, positive, linear relationship. Taller students tend to have longer arm spans.
Part (b): Classify body shapes (square, tall rectangle, short rectangle) based on arm span vs. height.
- (i) Helpful graph: Graph 2 (line $y=x$) is helpful. It visually separates the categories: points on the line are square (arm span = height), below the line are tall rectangles (arm span < height), and above the line are short rectangles (arm span > height).
- (ii) Classification table: Count points relative to the $y=x$ line.
  - Square (on line): 3 seniors
  - Tall Rectangle (below line): 4 seniors
  - Short Rectangle (above line): 5 seniors
Part (c): Calculate the predicted arm span for a height of 61 inches using the regression line $y = 11.74 + 0.8247x$.
- Calculation: Predicted $y = 11.74 + 0.8247(61) \approx 11.74 + 50.3067 \approx 62.05$ inches.

Question 6: Sampling Methods and Distributions

Scenario: A facility produces 100,000 tortillas/day on two production lines (A: mean diameter 5.9 in, B: mean diameter 6.1 in). A sample of 200 tortillas is taken to estimate the mean diameter.
Method 1: Random sample of 200 from all 200,000 tortillas.
Method 2: Randomly select one line, then sample 200 from that line's 100,000 tortillas.
Part (a): Will Method 2 be representative? Explain.
- Answer: No. Because the mean diameters of the two lines differ (5.9 vs 6.1), a sample from only one line will not represent the overall population mean.
Part (b): Identify the sampling method used to obtain a bimodal histogram.
- Answer: Method 1. A bimodal histogram suggests data from two different distributions (the two production lines), which would occur if sampling from both lines. Method 2 would yield a unimodal histogram centered near either 5.9 or 6.1.
Part (c): Which method results in less variability in the sample? Explain.
- Answer: Method 2. Sampling from a single production line results in less variability than sampling from two lines with different means.
Part (d): Describe the sampling distribution of the sample mean diameter (Method 1, sample size 200). Population mean = 6 in, population SD = 0.11 in.
- Description: Approximately normal (by Central Limit Theorem), with mean $\mu_{\bar{x}} = 6$ inches and standard deviation $\sigma_{\bar{x}} = \sigma/\sqrt{n} = 0.11/\sqrt{200} \approx 0.0078$ inches.
Part (e): Which method results in less variability in the distribution of 365 daily sample means? Explain.
- Answer: Method 1. Method 2 would produce sample means centered around 5.9 and 6.1, leading to greater overall variability across the year. Method 1's sample means will cluster tightly around the overall mean of 6.
Part (f): Which method is more likely to produce a sample mean close to 6 inches on the day of sampling? Explain.
- Answer: Method 1. Although both are unbiased estimators, Method 1 has less variability in its sample means, making it more likely that a single sample mean will be close to the population mean of 6 inches.

Section I: Multiple-Choice Questions

This section contains 40 multiple-choice questions covering a broad range of AP Statistics topics. The provided text includes the answer key for some questions.

Key Topics Covered (based on question examples):
- Measures of central tendency (median, mean, percentiles).
- Sampling methods (stratified random sampling, convenience sampling, simple random sample).
- Standardized scores (z-scores) and their invariance under linear transformations.
- Sources of bias (nonresponse bias, sampling bias).
- Estimating population parameters from sample statistics.
- Describing distribution shapes (skewness, normality).
- Interpreting boxplots (median, quartiles, spread, outliers).
- Confidence intervals and margin of error interpretation.
- Probability calculations (normal distribution, binomial distribution).
- Hypothesis testing (two-sample t-test, matched-pairs t-test, z-test for proportion, chi-square goodness-of-fit).
- Regression analysis (correlation, slope, residuals).
- Sampling distributions (mean, proportion).
- Properties of estimators (bias, variability).

Statistical Formulas and Tables

The document includes reference pages for:

Descriptive Statistics Formulas: Standard deviation, pooled standard deviation, correlation coefficient.
Probability Formulas: Union of events, binomial distribution, properties of sample means.
Inferential Statistics Formulas: Standardized test statistic, confidence intervals, difference of means/proportions, chi-square test statistic.
Statistical Tables: Standard normal probabilities (Table A), t-distribution critical values (Table B), $\chi^2$ critical values (Table C).

Additional Sections

Multiple-Choice Answer Key: Provides correct answers for selected multiple-choice questions.
Free-Response Scoring Guidelines: Detailed rubrics for grading the free-response questions, outlining criteria for "Essentially Correct," "Partially Correct," and "Incorrect" responses for each part.
Scoring Worksheet and Conversion Chart: Tools for calculating composite scores and converting them to AP grades.
Exam Instructions: Administrative details for proctors and students regarding exam administration, materials, and procedures.

Here's a summary of the provided AP Statistics content, organized by question:

AP Statistics Practice Exam - Section I & II Summary

This document contains a series of AP Statistics practice questions covering various topics, including descriptive statistics, probability, sampling, experimental design, and inferential statistics. It also includes scoring guidelines and performance data from the 2018 exam.

Section I: Multiple Choice Questions

1. Histogram of Meal Prices

Topic: Descriptive Statistics (Interquartile Range)
Content: A histogram shows the distribution of prices for 100 meals. The question asks to estimate the interquartile range (IQR) based on the histogram.

2. Scale Calibration and Correlation

Topic: Descriptive Statistics (Correlation)
Content: A scale consistently displays a mass 0.75 kg less than the actual mass. The question asks for the correlation between actual masses and displayed masses.
Key Insight: A linear transformation (adding or subtracting a constant) does not change the correlation.

3. Animal Weights and Classification

Topic: Data Visualization
Content: A table shows the number of cats and dogs classified by weight (healthy, underweight, overweight). The question asks for the most appropriate graph to show the relationship between animal type and weight classification.
Appropriate Graph: Segmented bar chart or side-by-side bar chart.

4. Stratified Random Sample

Topic: Sampling Methods
Content: A program aims to estimate the proportion of students who major in math/science after completing it. The question asks to identify the sampling method that describes a stratified random sample.
Definition: Stratified random sampling involves dividing the population into subgroups (strata) and then taking a random sample from each subgroup.

5. Sampling Distribution of the Sample Mean

Topic: Sampling Distributions
Content: A normal curve represents the sampling distribution of the sample mean ($\bar{x}$) for a sample size $n=25$ from a population with standard deviation $\sigma$. The question asks to estimate the population standard deviation ($\sigma$).
Formula: The standard deviation of the sampling distribution of the mean is $\sigma_{\bar{x}} = \frac{\sigma}{\sqrt{n}}$.

6. Margin of Error in Confidence Intervals

Topic: Statistical Inference (Confidence Intervals)
Content: Two random samples (A and B) from the same population are used to create 95% confidence intervals for the population mean. Sample A has a larger margin of error than Sample B. The question asks for a possible explanation.
Key Insight: The margin of error is directly proportional to the sample size ($n$) and the sample standard deviation ($s$). A larger margin of error typically implies a smaller sample size or a larger standard deviation.

7. Statistical Measures and Currency Conversion

Topic: Descriptive Statistics (Transformations)
Content: Annual sales data in Canadian dollars (CAD) are converted to US dollars (USD) using a fixed exchange rate (1 CAD = 0.75 USD). The question asks which statistic will have the same value in both currencies.
Key Insight: Measures of spread (standard deviation, IQR) and standardized scores are invariant to linear transformations that involve multiplication or division by a constant. Measures of central tendency (mean, median) and minimum/maximum will change.

8. Chi-Square Test Contribution

Topic: Statistical Inference (Chi-Square Test)
Content: A restaurant manager tracks weekly dinner orders. A sample of 200 orders is taken to compare current distribution to last year's distribution. The question asks for which dinner type contributes the most to the chi-square test statistic.
Formula: The contribution of each category to the chi-square statistic is $\frac{(\text{observed} - \text{expected})^2}{\text{expected}}$. A larger contribution comes from a larger squared difference between observed and expected counts, relative to the expected count.

9. Standard Deviation of a Linear Combination of Independent Random Variables

Topic: Probability (Random Variables)
Content: $X$ is the weight of fleece from Northern Farm (mean 14.1, std dev 1.3). $Y$ is the weight from Western Farm (mean 6.7, std dev 0.5). $W$ is the total weight from 10 sheep from Northern Farm and 15 from Western Farm. The question asks for the standard deviation of $W$.
Formula: For independent random variables, $Var(aX + bY) = a^2 Var(X) + b^2 Var(Y)$. The standard deviation is the square root of the variance.

10. Probability of Union of Events

Topic: Probability
Content: 36.5% of households own dogs, 30.4% own cats. The question asks for the probability a randomly selected household owns a dog OR a cat.
Formula: $P(A \cup B) = P(A) + P(B) - P(A \cap B)$.
Missing Information: The probability of owning both a dog and a cat ($P(A \cap B)$) is not provided, making the question unanswerable with the given information.

Section II: Free-Response Questions

Question 1: Completion Times (Histograms, Combined Distributions, Sampling Distributions)

Topic: Descriptive Statistics, Probability, Sampling Distributions
Part (a): Identify which histogram (I or II) represents Group R (students with no spreadsheet experience) and justify the choice.
- Reasoning: Group R, having no prior experience, is expected to take longer to complete the task. Histogram II shows generally higher completion times, making it more likely to represent Group R.
Part (b): Describe the shape of a histogram created by combining the data from both groups.
- Description: The combined distribution would likely be bimodal, with peaks corresponding to the typical completion times of Group S (shorter times) and Group R (longer times).
Part (c): Describe the sampling distribution of the sample mean completion time for random samples of 50 students from the state population (mean=70, std dev=26.5).
- Description: By the Central Limit Theorem (CLT), since the sample size ($n=50$) is large, the sampling distribution of the sample mean ($\bar{x}$) will be approximately normal.
- Parameters:
  - Mean of the sampling distribution ($\mu_{\bar{x}}$) = Population mean ($\mu$) = 70 minutes.
  - Standard deviation of the sampling distribution ($\sigma_{\bar{x}}$) = $\frac{\sigma}{\sqrt{n}} = \frac{26.5}{\sqrt{50}} \approx 3.75$ minutes.

Question 2: Running Shoe Experiment (Experimental Design)

Topic: Experimental Design (Blocking, Randomization, Replication)
Part (a): Statistical advantage of blocking by runner classification (professional vs. recreational).
- Advantage: Blocking accounts for known variation in running speed between professional and recreational runners. This reduces variability within blocks, making it easier to detect a significant difference between the shoe types.
Part (b): Importance of randomizing shoe assignment instead of allowing choice.
- Importance: Random assignment prevents confounding variables. If runners chose their shoes, there might be systematic differences between the groups (e.g., faster runners choosing a specific shoe type), which could distort the results. Randomization ensures groups are comparable on average.
Part (c): How replication is addressed and its benefit.
- Replication: Multiple runners are assigned to each shoe type within each block (e.g., multiple professional runners get shoe A, multiple get shoe B).
- Benefit: Replication allows for the estimation of natural variability in running times within each treatment group. This estimate is crucial for statistical inference (hypothesis testing, confidence intervals) to determine if observed differences are statistically significant or due to chance.

Question 3: STEM Internships (Probability, Sampling Methods)

Topic: Probability, Sampling Methods
Part (a): Probability that a randomly selected woman does not meet the age requirement (age < 20).
- Calculation: Sum probabilities for ages 17, 18, and 19: $0.005 + 0.107 + 0.111 = 0.223$.
Part (b): Probability that at least 30% of 100 women in a simple random sample do not meet the age requirement.
- Model: Let $X$ be the number of women who do not meet the age requirement. $X \sim \text{Binomial}(n=100, p=0.223)$. We need $P(X \ge 30)$.
- Calculation: Using a normal approximation or exact binomial calculation, find $P(X \ge 30)$. (The provided scoring guideline shows $P(X \ge 30) \approx 0.0453$).
Part (c): Compare likelihood of selection for a woman not meeting the age requirement using stratified vs. simple random sampling.
- Comparison: A woman who does not meet the age requirement is more likely to be selected with stratified sampling.
- Justification: The stratified sample specifically selects 30 women (30% of the sample) who do not meet the age requirement. In a simple random sample, the expected proportion is only 22.3%, making it less likely to capture this specific subgroup at a 30% rate.

Question 4: Activity Tracker Steps (Confidence Intervals)

Topic: Statistical Inference (Confidence Intervals)
Part (a): Construct and interpret a 99% confidence interval for the mean number of steps.
- Procedure: One-sample t-interval.
- Conditions: Random sample (stated), large sample size ($n=61 > 30$).
- Calculation:
  - Sample mean ($\bar{x}$) = 9,797 steps.
  - Sample standard deviation ($s$) = 2,313 steps.
  - Degrees of freedom ($df$) = $n-1 = 60$.
  - Critical t-value ($t^*$) for 99% confidence and 60 df $\approx 2.660$.
  - Margin of Error (ME) = $t^* \times \frac{s}{\sqrt{n}} = 2.660 \times \frac{2313}{\sqrt{61}} \approx 788.5$.
  - Confidence Interval: $\bar{x} \pm ME = 9797 \pm 788.5 = (8908.5, 10585.5)$.
- Interpretation: "We are 99% confident that the true mean number of steps taken on a typical workday by people in New York City who wear activity trackers is between 8,909 and 10,586 steps."
Part (b): Appropriateness of using the CI to investigate if 8,500 steps is unusual for an individual.
- Answer: No, it is not appropriate.
- Explanation: A confidence interval estimates the population mean, not individual values. It provides a range for the average number of steps, not the range of steps for any single person. To investigate an individual value, one might use a z-score or construct a prediction interval.

Question 5: Race Times (Scatterplots, Correlation, Extrapolation)

Topic: Descriptive Statistics (Scatterplots, Correlation, Regression)
Part (a): Identify the winner and approximate total race time.
- Method: The winner has the minimum total time (Reaction Time + Running Time). Locate the point on the scatterplot closest to the line $y = -x + C$ where $C$ is minimized. This corresponds to the point with the smallest sum of x and y coordinates.
- Approximation: Based on the graph, the point with approximately (Reaction Time = 0.152, Running Time = 9.61) represents the winner, with a total time of approximately $0.152 + 9.61 = 9.762$ seconds.
Part (b): Reasonableness of assuming independence between reaction time and running time.
- Answer: No, it is not reasonable.
- Justification: The scatterplot shows a strong positive linear association between reaction time and running time. This indicates they are not independent; as reaction time increases, running time tends to increase.
Part (c): Appropriateness of predicting running time for a reaction time of 0.30 seconds using the regression model.
- Explanation: It is likely not appropriate because 0.30 seconds is far outside the range of observed reaction times (which appear to go up to about 0.202 seconds). This is extrapolation, and the linear relationship observed within the data range may not hold for values outside that range.

Question 6: Lightbulb Defects (Hypothesis Testing, Sampling Distributions, Chebyshev's Inequality)

Topic: Statistical Inference, Probability
Part (a): Condition for normal distribution of $\hat{p}_x$ (proportion of defective brand-name bulbs).
- Condition: The number of successes ($np$) and failures ($n(1-p)$) should be at least 10.
- Check: Observed: $n=400$, $\hat{p}_x = 44/400 = 0.11$. Expected successes: $n \times p_0 = 400 \times 0.10 = 40$. Expected failures: $n \times (1-p_0) = 400 \times 0.90 = 360$. Both 40 and 360 are $\ge 10$. Yes, it is reasonable.
Part (b): Show calculation of standard error for $\hat{p}_x$.
- Formula: $SE(\hat{p}_x) = \sqrt{\frac{p_0(1-p_0)}{n}} = \sqrt{\frac{0.10(1-0.10)}{400}} = \sqrt{\frac{0.10(0.90)}{400}} = \sqrt{\frac{0.09}{400}} = \sqrt{0.000225} = 0.015$. (Note: The provided SE was 0.0156, likely using $\hat{p}_x$ instead of $p_0$ or a slightly different calculation).
Part (c): Number of standard errors the observed $\hat{p}_x$ is from 0.10.
- Calculation: $z = \frac{\hat{p}_x - p_0}{SE(\hat{p}_x)} = \frac{0.11 - 0.10}{0.0156} \approx 0.64$ standard errors.
Part (d): Calculate $\hat{D}$ (estimate of average proportion) and its standard error ($SE(\hat{D})$).
- Calculation:
  - $\hat{p}_x = 44/400 = 0.11$.
  - $\hat{p}_y = 104/400 = 0.26$.
  - $\hat{D} = \frac{\hat{p}_x + \hat{p}_y}{2} = \frac{0.11 + 0.26}{2} = \frac{0.37}{2} = 0.185$.
  - $SE(\hat{D}) = \sqrt{\frac{SE(\hat{p}_x)^2 + SE(\hat{p}_y)^2}{4}} = \sqrt{\frac{(0.0156)^2 + (0.0219)^2}{4}} \approx \sqrt{\frac{0.00024336 + 0.00047961}{4}} \approx \sqrt{\frac{0.00072297}{4}} \approx \sqrt{0.0001807} \approx 0.0134$.
Part (e): Calculate the test statistic $W = \hat{D} - 0.10$.
- Calculation: $W = 0.185 - 0.10 = 0.085$.
Part (f): Use Chebyshev's inequality to decide if there's evidence $D > 0.10$ at $\alpha = 0.05$.
- Chebyshev's Inequality: $P(|X - \mu| \ge k\sigma) \le \frac{1}{k^2}$.
- Test Statistic Value: The question implies $W$ is the test statistic, but the formula provided is $W = \hat{D} - 0.10$. The test statistic for hypothesis testing is usually standardized. Assuming the question meant to use the standardized value: $z = \frac{\hat{D} - 0.10}{SE(\hat{D})} = \frac{0.185 - 0.10}{0.0134} \approx 6.34$.
- Applying Chebyshev: If $k=6.34$, $P(| \text{statistic} - \text{mean} | \ge 6.34 \times SE) \le \frac{1}{6.34^2} \approx \frac{1}{40.2} \approx 0.0249$.
- Conclusion: The observed value is 6.34 standard errors above the mean. The probability of observing a value this extreme or more extreme is at most 0.0249. Since $0.0249 < 0.05$, we reject the null hypothesis. There is sufficient statistical evidence at the 0.05 level to conclude that the average proportion of defective lightbulbs is greater than 0.10.

This summary covers the main points and calculations presented in the provided text. The specific answer choices for the multiple-choice questions are not included as they require the original question context. The free-response sections are summarized based on the tasks and scoring guidelines provided.

Here's a summary of the provided AP Statistics practice exam questions, organized by topic:

AP Statistics Practice Exam Summary

This document contains a series of multiple-choice and free-response questions from an AP Statistics practice exam, covering various statistical concepts and their applications.

Section I: Multiple Choice Questions

This section consists of 40 multiple-choice questions designed to assess understanding of core statistical concepts. The questions cover a range of topics, including:

Descriptive Statistics:
- Identifying and describing distributions (shape, center, spread, outliers) from dotplots, histograms, and stemplots (Questions 1, 4, 24).
- Interpreting boxplots for comparison (Question 7).
- Calculating and interpreting correlation and regression (Question 21).
Probability and Distributions:
- Calculating conditional probabilities (Question 2).
- Understanding and applying normal distributions (Questions 6, 10, 17, 22).
- Understanding and applying binomial distributions (Questions 18, 28).
- Understanding sampling distributions of sample means and proportions (Questions 13, 18, 22).
- Understanding geometric distributions (implied in Question 28).
Statistical Inference:
- Constructing and interpreting confidence intervals for proportions and means (Questions 14, 31, 32, 40).
- Constructing and interpreting confidence intervals for regression slopes (Question 33).
- Performing and interpreting hypothesis tests for proportions and means (Questions 15, 25, 34, 36).
- Performing and interpreting hypothesis tests for regression slopes (Question 35).
- Performing and interpreting chi-square tests (Questions 23, 39).
- Understanding Type II errors (Question 29).
- Understanding the impact of sample size on sampling distributions (Question 31).
Study Design:
- Identifying types of sampling methods (convenience, simple random, stratified, systematic) (Question 12).
- Identifying features of experimental design (replication, treatments, experimental units, response variable, randomized block design, matched-pairs design) (Questions 16, 27).
- Generalizing results from a study to a population (Question 19).
- Distinguishing between observational studies and experiments (Question 8).
Data Analysis and Interpretation:
- Interpreting statistical output (e.g., regression output) (Question 33).
- Interpreting p-values (Questions 15, 25, 34).
- Understanding the relationship between correlation and regression (Question 21).
- Calculating contributions to a chi-square statistic (Question 39).
- Calculating standard error for proportions (Question 38).
- Calculating mean and standard deviation of random variables (Question 17, 37).

Section II: Free Response Questions

This section consists of 6 free-response questions, divided into Part A (5 questions) and Part B (1 question), assessing the ability to apply statistical concepts in more depth.

Part A (Questions 1-5):

Describing Distributions and Identifying Outliers:
- Requires describing a procedure for identifying potential outliers (e.g., using IQR) and applying it to data presented in a stemplot.
- Also requires describing the shape, center, and spread of a distribution based on a stemplot.
Regression Inference:
- Involves constructing and interpreting a confidence interval for the slope of a least-squares regression line.
- Requires determining if the confidence interval contradicts a given belief about the relationship.
Comparing Distributions (Histograms):
- Involves extracting counts from histograms to answer specific questions about the data.
- Requires comparing the shapes, centers, and spreads of two distributions presented as histograms.
- Requires estimating the median of a combined distribution based on histograms.
Experimental Design (Randomized Block Design):
- Asks to identify conditions under which a randomized block design is superior to a completely randomized design.
- Requires identifying blocks and describing the random assignment of treatments within blocks.
Binomial and Sampling Distributions:
- Requires calculating binomial probabilities for "at least" and "first occurrence" scenarios.
- Assesses understanding of the conditions for approximating a binomial distribution with a normal distribution and justifying the approximation.

Part B (Question 6):

Nonparametric Tests (Wilcoxon Rank Sum Test):
- Requires explaining why a two-sample t-test is inappropriate for small sample sizes with potential outliers.
- Requires justifying the use of medians over means in such cases.
- Involves ranking combined data and calculating the sum of ranks.
- Requires calculating a test statistic (W) for the Wilcoxon Rank Sum Test.
- Involves using a provided sampling distribution graph to find a p-value and make a conclusion in context at a given significance level.

Appendices and Supporting Materials

Formula Sheet: Provides formulas for descriptive statistics, probability, binomial distributions, and inferential statistics (standardized test statistics, confidence intervals, standard deviations of statistics).
Statistical Tables: Includes standard normal (Table A), t-distribution (Table B), and chi-square (Table C) critical values.
Scoring Guidelines: Detailed rubrics for the free-response questions, outlining criteria for "Essentially Correct," "Partially Correct," and "Incorrect" responses for each part. These guidelines also provide examples of acceptable answers and common errors.
Exam Administration Instructions: Notes for proctors on administering the exam, including timing, student instructions, and material collection procedures.
Question Descriptors and Performance Data: Lists the statistical concepts assessed by each multiple-choice question and provides performance data for the 2019 exam.

Here's a summary of the provided content, organized into sections:

Summary of AP Statistics Exam Excerpts

This document contains excerpts from an AP Statistics exam, including multiple-choice questions and free-response sections, along with scoring guidelines and administrative instructions. The content covers various statistical concepts and their application.

Section I: Multiple-Choice Questions

This section presents a series of multiple-choice questions designed to test understanding of statistical concepts.

Question 23: Focuses on the relationship between sample size and the margin of error for a one-sample z-interval for a proportion. It asks how the margin of error changes when the sample size increases, assuming equal point estimates.
Question 24: Describes a study investigating the impact of a nutritional supplement on reaction time, with participants grouped by athletic ability. It requires identifying the study design.
- Study Design: Randomized block design, where athletic ability is the blocking variable.
Question 25: Involves calculating the mean cost per customer based on a relative frequency distribution of the number of tickets purchased and the cost per ticket.
Question 26: Pertains to expected counts in a survey comparing male and female teens' brushing habits, assuming no difference in population proportions.
Question 27: Discusses the interpretation of a p-value from a significance test for the slope of a regression line, relating it to the association between lean body mass and maximal oxygen uptake.
Question 28: Compares two confidence intervals for a proportion constructed at different times (winter and summer) from samples of the same size, given different sample proportions. It asks how the width and point estimate of the summer interval compare to the winter interval.
Question 29: Uses boxplots to compare the distribution of scores for golfers in two different tournaments. It asks which statement must be true based on the boxplots.
Question 30: Describes the distribution of sample means obtained from repeated sampling from a population with a right-skewed distribution. It requires understanding the Central Limit Theorem's effect on the sampling distribution of the mean.
Question 31: Involves analyzing residuals from a least-squares regression line to assess model fit. It asks what a plot of residuals versus the dose might reveal.
Question 32: A probability problem involving a binomial scenario (or geometric if considering the first success), calculating the probability of at least one success in a fixed number of trials.
Question 33: Asks to identify the correct formula for a confidence interval for the slope of a least-squares regression line, given summary statistics and a confidence level.
Question 34: Calculates the expected value and standard deviation of a player's score in a card game, based on the expected values and standard deviations of points earned and minutes played, assuming independence.
Question 35: Critiques the conclusions drawn from a study about a spider-fear program, focusing on potential biases due to volunteer participation and self-selection.
Question 36: Defines the "power of the test" in the context of a police officer issuing speeding tickets.
Question 37: Evaluates the appropriateness of a conclusion drawn from an experiment comparing fertilizers on green bean growth, considering generalization and study design.
Question 38: Explores how changes in sample size, sample mean difference, and sample standard deviation affect the p-value of a hypothesis test.
Question 39: Tests the interpretation of a 95 percent confidence level for a population proportion.
Question 40: Assesses the conclusion of a hypothesis test comparing two population proportions, based on the p-value.

Section II: Free-Response Questions

This section contains free-response questions that require detailed explanations, calculations, and interpretations.

Part A: Questions 1-5

Question 1: Involves interpreting population pyramids to compare age distributions between two countries and identifying historical birth rate changes.
- (a) Compares the proportion of females aged 60+ in Country A vs. Country B.
- (b) Identifies which country experienced specific birth rate increases based on age group distributions.
- (c) Determines the median age group for the male population in Country A.
Question 2: Investigates whether playing video games improves visual search ability.
- (a) Describes a method for randomly assigning participants to treatment groups.
- (b) States the appropriate statistical test and hypotheses to compare mean improvement times between two game types.
- (c) Suggests changes to increase the power of a test when no significant difference was found.
Question 3: Deals with probability calculations for a tennis match where sets are played sequentially.
- (a) Lists all possible sequences of set wins for each player to win the match.
- (b) Calculates the probability of Player V winning the match, given specific conditional probabilities.
- (c) Calculates a conditional probability: the probability of a 3-set match given Player V wins.
- (d) Calculates the expected number of sets played in a match.
Question 4: Analyzes data on the ages of maternal grandparents.
- (a) Constructs and interprets a 95 percent confidence interval for the mean difference in age (grandmother - grandfather).
- (b) Provides a point estimate for the mean difference in age without regard to which grandparent is older (i.e., absolute difference).
Question 5: Examines the proportion of new cars experiencing mechanical problems.
- (a) Describes a plan to obtain a simple random sample of customers.
- (b) Explains why a conditional proportion (e.g., proportion of problems given a car had any problem) should not be used to estimate the population proportion of cars with a specific problem (power door locks).
- (c) Calculates a point estimate for the number of cars sold that experienced a problem with the power door locks.

Part B: Question 6

Question 6: Focuses on identifying outliers and comparing methods of random assignment.
- (a) Determines if a specific value (7.5 miles) is an outlier using the 1.5 * IQR rule.
- (b) Derives the mean and standard deviation of the difference between two independent normal random variables ($Y - X$).
- (c) Estimates the mean and standard deviation of the difference $Y - X$ using sample data.
- (d) Calculates a test statistic for a hypothesis test comparing the means of two populations, where one value is suspected to be an outlier.
- (e) Draws conclusions about population means and the outlier status based on a given p-value.

Administrative and Scoring Information

The document also includes:

Exam Instructions: Details for proctors and students regarding exam administration, including timing, materials, and procedures for both multiple-choice and free-response sections.
Formulas and Tables: Standard statistical formulas and critical values (z, t, $\chi^2$) provided for student use.
Answer Key: Correct answers for the multiple-choice questions.
Scoring Guidelines: Detailed rubrics for grading each free-response question, outlining criteria for "Essentially Correct (E)," "Partially Correct (P)," and "Incorrect (I)" responses for each part. These guidelines explain how points are awarded based on statistical concepts, calculations, interpretations, and justifications.
Scoring Worksheet and Conversion Chart: Information on how raw scores are converted into AP scores (1-5).
Performance Data: Statistics on how students performed on individual multiple-choice questions.

This summary provides a structured overview of the content presented in the AP Statistics exam excerpts, highlighting the key statistical concepts and tasks involved.

AP® Statistics 2023 Scoring Guidelines Summary

This document outlines the scoring criteria for the AP® Statistics exam, detailing how responses are evaluated for various question types, including hypothesis testing, confidence intervals, data exploration, experimental design, probability, and investigative tasks.

Section 1: Hypothesis Testing (Paired t-test Example)

This section details the scoring for a paired t-test, focusing on checking conditions, calculating statistics, and interpreting results.

Key Components for "Essentially Correct" (E) Scoring:

Independence Condition: Verified by referring to random assignment in the experiment.
- Note: Random sampling of patients does not satisfy this condition.
Normality Condition: Supported by the distribution of sample differences being approximately normal (e.g., symmetric with no outliers in a boxplot), even with a small sample size (n=19).

Test Statistic: Correctly calculated and reported value consistent with the named test (e.g., t-statistic).

Example Calculation:

t = (mean difference - hypothesized mean) / (standard deviation of differences / sqrt(n))
t = (0 - 0) / (1.789 / sqrt(19)) ≈ 3.138

p-value: Correctly reported and consistent with the stated alternative hypothesis and test statistic.
- Example p-value: P(t > 3.138) with 18 degrees of freedom ≈ 0.0028.

Scoring Breakdown:

Essentially Correct (E): Satisfies all four components.
Partially Correct (P): Satisfies two or three components.
Incorrect (I): Satisfies fewer than two components.

Important Clarifications:

Degrees of freedom discussion is considered extraneous.
If a z-test is incorrectly identified, specific scoring adjustments apply to the test statistic and p-value.
Comparing the test statistic to a critical value instead of computing a p-value can satisfy component 4 if done correctly.
A two-tailed alternative hypothesis requires a p-value consistent with that hypothesis.
Errors in supporting work for the t-statistic may be overlooked if the final value is correct.
If no test statistic is reported, component 4 can still be satisfied if the p-value is correct.

Confidence Interval Approach:

This approach can also be used to satisfy components 3 and 4.
Correctly calculated confidence intervals (e.g., one-sided 95% or two-sided 90%) can satisfy component 3.
Correct interpretation of the interval (e.g., comparing zero to the interval) can satisfy component 4.
Specific interval calculations are provided for different scenarios (e.g., ua vs. ud).

Section 2: Inference Conclusion

This section focuses on the interpretation of the hypothesis test results.

Key Components for "Essentially Correct" (E) Scoring:

Comparison and Decision: Correctly compares the p-value to the significance level (alpha) and makes a correct decision about the null hypothesis (reject or fail to reject).
- Example: p-value (0.0028) < alpha (0.05), so reject the null hypothesis.
Conclusion in Context: States a conclusion in context, consistent with the alternative hypothesis, using non-deterministic language.
- Example: "The data provide convincing statistical evidence that for patients similar to those in the study, the true mean difference (placebo minus omega-3) in irritability scores... is greater than zero." This suggests omega-3 helps reduce irritability.

Scoring Breakdown:

Essentially Correct (E): Satisfies both components.
Partially Correct (P): Satisfies only one component.
Incorrect (I): Satisfies neither component.

Important Clarifications:

The p-value must be clearly identified when compared to alpha.
Comparing the test statistic to a critical value can satisfy component 1.
An explicit decision is not required if implied in the conclusion.
Context in component 2 must include the parameter, population/sampling units, and response variable.
Non-deterministic language (e.g., "evidence to suggest") is required; deterministic language (e.g., "proves") is incorrect.
Responses based on incorrect p-values or test statistics can still satisfy components if the comparison and conclusion are consistent.
Confidence interval approach can be used, comparing zero to the interval.

Section 3: Question Scoring Overview

This section provides a general rubric for scoring multi-part questions based on the "Essentially Correct" (E), "Partially Correct" (P), and "Incorrect" (I) ratings of each part.

4 points: Three parts E.
3 points: Two parts E and one part P.
2 points: Two parts E and no part P, OR one part E and one or two parts P, OR three parts P.
1 point: One part E and no part P, OR no part E and two parts P.
0 points: No part E and one part P, OR no parts E or P.

Question 5: Focus on Exploring Data (Example: Tule Elk)

This question assesses the ability to describe relationships in data.

(a) Scatterplot Description:

Essentially Correct (E): Requires at least three components from (1-4) plus context (5).
1. Direction: Positive/increasing.
2. Strength: Strong.
3. Form: Linear/approximately linear.
4. Unusual Features: No significant outliers or deviations.
5. Context: Association between chest circumference and weight of tule elk.
Partially Correct (P): Satisfies fewer components.
Incorrect (I): Does not meet E or P criteria.

(b) Prediction and Residual:

Essentially Correct (E): Requires correct calculation of predicted weight (b-i) and residual (b-ii) with work shown.
- Example (b-i): Predicted weight = -350.3 + 3.7455(145.9) ≈ 196.17 kg.
- Example (b-ii): Residual = Actual - Predicted = 204.3 - 196.17 ≈ 8.13 kg.
Partially Correct (P): Satisfies only one component or both correct values with no work.
Incorrect (I): Does not meet E or P criteria.

(c) Interpretation of Slope:

Essentially Correct (E): Requires identifying the slope value, interpreting its meaning in context (increase in weight per unit increase in circumference), and using non-deterministic language for prediction.
1. Value: 3.7455.
2. Interpretation: Increase in kg per cm increase in circumference.
3. Prediction: Uses terms like "predicted," "estimated," "expected."
Partially Correct (P): Satisfies two components.
Incorrect (I): Does not meet E or P criteria.
- Note: Specific rounding of the slope is acceptable. Units must be correct.

Question 1: Focus on Exploring Data (Example: Dissolved Oxygen)

This question assesses the description of a distribution from a histogram and boxplot.

(a) Histogram Description:

Essentially Correct (E): Requires context (1) AND at least three of shape (2), center (3), spread (4), and unusual features (5).
1. Context: Dissolved oxygen concentration.
2. Shape: Skewed left.
3. Center: Median between 11-12 mg/l.
4. Spread: Mention of range, IQR, or values between 2-14 mg/l.
5. Unusual Features: Potential outliers or a gap.
Partially Correct (P): Satisfies context and two other components, OR three other components without context.
Incorrect (I): Does not meet E or P criteria.
- Note: Describing the distribution as normal is incorrect for shape. Center and spread descriptions require approximate language and appropriate values.

(b) Boxplot Construction:

Essentially Correct (E): Requires five specific components related to boxplot elements:
1. Box starts at Q1.
2. Box ends at Q3.
3. Median located within the box.
4. Whiskers extend to the minimum.
5. Whiskers extend to the maximum.
Partially Correct (P): Satisfies three components.
Incorrect (I): Does not meet E or P criteria.
- Note: Specific placement of elements relative to tick marks is important.

(c) Comparing Distributions:

Essentially Correct (E): Requires context statement (1) AND at least two of center comparison (2), shape comparison (3), and spread comparison (4).
1. Context Statement: Colder streams are healthier.
2. Center: Compare medians (colder > warmer).
3. Shape: Compare skewness (colder left, warmer right).
4. Spread: Compare IQR or range (similar spread).
Partially Correct (P): Satisfies context and one other component.
Incorrect (I): Does not meet E or P criteria.
- Note: Comparisons must be direct. Outliers don't need to be compared.

Question 2: Focus on Sampling and Experimental Design

This section assesses understanding of experimental design principles.

(a) Identifying Experimental Components:

Essentially Correct (E): Requires identifying:
1. Experimental Units: Driveways (or new homes needing driveways).
2. Treatments: Concrete with fibers vs. without fibers.
3. Response Variable: Rating of crack severity (0-10 scale).
Partially Correct (P): Satisfies two components.
Incorrect (I): Does not meet E or P criteria.

(b) Random Assignment Procedure:

Essentially Correct (E): Requires:
1. Random Process: Correct use of a random number generator (or equivalent) to assign units to treatments, ensuring all assignments are equally likely.
2. Equal Numbers: Assignment results in an equal number of units (30) per treatment.
Partially Correct (P): Satisfies only one component.
Incorrect (I): Does not meet E or P criteria.
- Note: Detailed examples of valid random assignment methods (slips of paper, urns, spinners) are provided. Processes without replacement or with clear stopping rules are crucial.

(c) Causation Conclusion:

Essentially Correct (E): Requires:
1. Causation: Random assignment allows concluding that the treatment (concrete type) caused the response (crack severity).
2. Context: Explanation is in the context of the problem (fibers, concrete, cracks).
Partially Correct (P): Satisfies only causation OR indicates random assignment limits confounding variables.
Incorrect (I): Does not meet E or P criteria.
- Note: Incorrect conclusions (generalization, bias reduction) lower the score.

Question 3: Focus on Probability and Sampling Distributions

This section covers basic probability calculations and expected value.

(a) Calculating Proportions:

Essentially Correct (E): Requires correctly calculating proportions (or percentages) for $1 prizes (0.68) and at least $10 prizes (0.12), with supporting work for both.
Partially Correct (P): Satisfies two components.
Incorrect (I): Does not meet E or P criteria.

(b) Conditional Probability:

Essentially Correct (E): Requires correctly calculating the conditional probability P(X=$100 | X≥$10) = 0.01 / 0.12 ≈ 0.0833, with work consistent with part (a-ii).
Partially Correct (P): Satisfies only one component.
Incorrect (I): Does not meet E or P criteria.

(c) Expected Value:

Essentially Correct (E): Requires stating the correct expected value ($4.68), showing appropriate work (sum of value * probability), and providing an interpretation that includes long-run average/mean in context.
- Example Work: E(X) = 1(0.68) + 5(0.2) + ... + 100(0.01) = $4.68.
Partially Correct (P): Satisfies calculation and work, plus one part of the interpretation.
Incorrect (I): Does not meet E or P criteria.
- Note: Calculator notation is not sufficient work.

(d) Currency Conversion:

Essentially Correct (E): Correctly converts the expected value to euros ($4.68 * 0.89 ≈ 4.17 euros), showing work and including units.
Partially Correct (P): Correct calculation but missing work or units.
Incorrect (I): Does not meet E or P criteria.

Question 4: Focus on Inference (Hypothesis Testing)

This section details the scoring for a full hypothesis test, broken into three sections.

Section 1: Hypotheses and Test Identification

Essentially Correct (E): Requires:
1. Test Identification: Paired t-test (name or formula).
2. Hypotheses Notation: Using a single mean symbol (e.g., µd).
3. Hypotheses Values: Correct null (e.g., µd = 0) and alternative (µd > 0).
4. Parameter Context: Defining the parameter (population mean difference) with sampling units and response variable.
Partially Correct (P): Satisfies three components.
Incorrect (I): Satisfies fewer than three components.
- Note: Hypotheses can be stated in words. Context can be provided later.

Section 2: Conditions, Test Statistic, and p-value

Essentially Correct (E):
1. Independence: Verified via random assignment.
2. Normality: Supported by distribution of differences (symmetric, no outliers).
3. Test Statistic: Correct calculation (t ≈ 3.138).
4. p-value: Correct calculation (p ≈ 0.0028).
Partially Correct (P): Satisfies two or three components.
Incorrect (I): Satisfies fewer than two components.
- Note: Specific scoring adjustments exist for z-test errors or confidence interval approaches.

Section 3: Conclusion

Essentially Correct (E):
1. Comparison & Decision: Correctly compares p-value to alpha and makes a decision (reject/fail to reject H0).
2. Contextual Conclusion: States conclusion in context, consistent with Ha, using non-deterministic language.
Partially Correct (P): Satisfies only one component.
Incorrect (I): Satisfies neither component.
- Note: Confidence interval approach can be used here as well.

Question 6: Investigative Task

This section assesses the ability to apply statistical concepts to a novel problem.

(a) Normal Distribution Probability:

Essentially Correct (E): Requires indicating a normal distribution (mean=300, sd=5), specifying the correct event (296 < X < 304), and providing the correct probability (≈ 0.5763).
Partially Correct (P): Satisfies two components.
Incorrect (I): Satisfies fewer than two components.
- Note: Various ways to indicate the normal distribution (graph, syntax, words, notation) are accepted.

(b) Sampling Distribution Probability and Interpretation:

Essentially Correct (E): Requires correct calculation of P(X̄ > 303) (≈ 0.198) with work shown (using N(300, 5/√2)), and a correct interpretation stating that this probability is not convincing evidence against the machine working properly, justified by the probability value.
Partially Correct (P): Satisfies at least three of the five components.
Incorrect (I): Satisfies fewer than three components.
- Note: Correctly calculating the standard deviation of the sample mean (σ/√n) is crucial. Interpretation must link probability to evidence.

(c) Distribution of Sample Range:

Essentially Correct (E): Requires describing the shape (skewed right), center (≈ 6 mg), and spread (values between 0-30 mg) for the sample range distribution (c-i), AND indicating that the mean and variation of the sample range increase as the population standard deviation increases (c-ii).
Partially Correct (P): Satisfies three components.
Incorrect (I): Satisfies fewer than three components.

(d) Unusualness of Sample Statistics:

Essentially Correct (E): Requires stating that a sample range of 10 mg is not unusual (d-i) and that Cleo's sample mean (303 mg) and range (10 mg) do not indicate the machine is faulty (d-ii). Both parts must be justified using probabilities or comparisons to standard deviations/typical values.
Partially Correct (P): Satisfies at least three components.
Incorrect (I): Satisfies fewer than three components.
- Note: Justifications should refer to probabilities calculated earlier or typical behavior of the distributions.

Summary of AP Statistics 2014 Free-Response Questions

This document contains the scoring guidelines and solutions for the free-response section of the 2014 AP Statistics Exam. It covers six questions, each designed to assess different statistical concepts and skills.

Question 1: Temperature Comparison and Variability

Main Idea: Comparing temperature patterns of two cities (Madison and Juneau) using a graph and assessing variability.
Part (a) Similarities and Differences:
- Similarities: Both cities exhibit a seasonal pattern with colder temperatures in winter (beginning/end of the year) and warmer temperatures in summer (middle of the year). Both show an increase in temperature from January to July and a decrease from July to December.
- Differences: Madison is generally warmer than Juneau for more months of the year. Madison's average monthly temperatures vary more than Juneau's.
Part (b) Standard Deviation:
- City with Greater Standard Deviation: Madison.
- Justification: The average monthly temperatures for Madison deviate more from its mean compared to Juneau's temperatures deviating from its mean. This greater variation in Madison's temperatures indicates a larger standard deviation.

Question 2: Experimental Design and Inference

Main Idea: Evaluating an experiment investigating the effect of black tea on heart health.
Part (a) Study Type:
- Type: Experiment.
- Explanation: Participants were randomly assigned to different treatment groups (tea without milk, tea with milk, hot water).
Part (b) Purpose of Hot Water Group:
- Reason: To serve as a placebo and control for the effect of simply drinking hot liquid. It helps determine if the benefits of tea are due to the tea itself or just the act of drinking a hot beverage. It also allows comparison of tea with milk versus tea without milk against a baseline.
Part (c) Generalizability:
- Conclusion: Not necessarily reasonable to generalize.
- Reason: The 24 participants were volunteers, who may not be representative of the larger population regarding the effect of tea on heart health.

Question 3: Hypothesis Testing for Treatment Effectiveness

Main Idea: Assessing the effectiveness of a new heart attack treatment using a two-sample t-test.
Hypotheses:
- $H_0: \mu_{new} = \mu_{standard}$ (or $\mu_{new} - \mu_{standard} = 0$)
- $H_a: \mu_{new} > \mu_{standard}$ (or $\mu_{new} - \mu_{standard} > 0$)
Test Procedure: Two-sample t-test.
Conditions:
- Random assignment to treatment groups (satisfied).
- Normal distribution of myocardial salvage index (MSI) values or large sample sizes. Sample sizes ($n=56$ for both groups) are large enough ($n>30$) for the t-test to be valid.
Test Statistic: $t \approx 0.69$ (calculated from sample means and standard deviations).
P-value: $p \approx 0.009$ (based on approximately 109 degrees of freedom).
Conclusion:
- Decision: Reject $H_0$.
- Interpretation: At the $\alpha = 0.05$ significance level, there is convincing statistical evidence that the new treatment produces a higher mean MSI than the standard care. This cause-and-effect conclusion is warranted due to the randomized experiment.

Question 4: Probability and Expected Value (Piñata)

Main Idea: Calculating probabilities and expected value related to a sequence of events with changing probabilities.
Part (a) Probability of Specific Sequence:
- $P(\text{No break on 1st, Break on 2nd}) = P(\text{No break on 1st}) \times P(\text{Break on 2nd | No break on 1st})$
- $P = (1 - 0.7) \times 0.8 = 0.3 \times 0.8 = 0.24$
Part (b) Probability Distribution of X (Number of Hits):
- $P(X=1) = 0.7$
- $P(X=2) = P(\text{No break on 1st}) \times P(\text{Break on 2nd | No break on 1st}) = 0.3 \times 0.8 = 0.24$
- $P(X=3) = P(\text{No break on 1st}) \times P(\text{No break on 2nd | No break on 1st}) \times P(\text{Break on 3rd | No break on 1st & 2nd})$
  - Probability of break on 3rd hit = $0.8 + 0.1 = 0.9$
  - $P(X=3) = 0.3 \times 0.2 \times 0.9 = 0.054$
- $P(X=4) = 1 - P(X=1) - P(X=2) - P(X=3) = 1 - 0.7 - 0.24 - 0.054 = 0.006$
Part (c) Expected Value of X:
- $E(X) = \sum x P(X=x)$
- $E(X) = 1(0.7) + 2(0.24) + 3(0.054) + 4(0.006) = 0.7 + 0.48 + 0.162 + 0.024 = 1.366$
- Interpretation: If Sophia were to break many piñatas under these conditions, the average number of hits required would be approximately 1.366 hits per piñata.

Question 5: Chi-Squared Test for Independence

Main Idea: Analyzing the association between gender and favorite sport using a chi-squared test.
Part (a) Association in the Sample:
- Summary: The graph reveals differences in sport preferences between males and females. Males tend to favor football and basketball more than females, while females tend to favor gymnastics and swimming more than males. Baseball preference is similar between genders.
Part (b) Statistical Test and Degrees of Freedom:
- Test: Chi-squared test for independence.
- Degrees of Freedom (df): $(R-1)(C-1)$, where R is the number of rows (genders) and C is the number of columns (sports).
  - $df = (2-1)(5-1) = 1 \times 4 = 4$.
Part (c) Conclusion from Test Statistic:
- Test Statistic: $\chi^2 = 10.77$.
- P-value: $p \approx 0.029$ (obtained from technology).
- Significance Level: $\alpha = 0.05$.
- Decision: Since $p < \alpha$ ($0.029 < 0.05$), reject $H_0$.
- Conclusion: There is convincing statistical evidence of an association between gender and the choice of favorite sport to watch for all students at the high school.

Question 6: Hypothesis Testing for Machine Malfunction (Penny Rolls)

Main Idea: Using a hypothesis test based on the weight of penny rolls to detect if a machine is underfilling.
Hypotheses:
- $H_0$: The machine is putting 50 pennies in each roll.
- $H_a$: The machine is putting fewer than 50 pennies in each roll.
Part (a) Type II Error:
- Description: A Type II error occurs if the inspector fails to reject $H_0$ (concludes the machine is functioning properly) when $H_a$ is actually true (the machine is putting fewer than 50 pennies in each roll).
- Consequence: The malfunctioning machine would continue to underfill rolls without being repaired, leading to continued production of rolls with fewer than 50 pennies.
Part (b) Distribution of W (Total Weight if 50 Pennies):
- Distribution: Approximately normal.
- Mean ($\mu_W$): $50 \times \mu_{\text{penny}} = 50 \times 2.5 \text{ grams} = 125 \text{ grams}$.
- Standard Deviation ($\sigma_W$): $\sqrt{50} \times \sigma_{\text{penny}} = \sqrt{50} \times 0.04 \text{ grams} \approx 0.2828 \text{ grams}$.
- Distribution: $W \sim N(125, 0.2828^2)$.
Part (c) Probability of Incorrect Conclusion (50 Pennies):
- Inspector's Rule: Conclude fewer than 50 pennies if weight $\le 124$ grams.
- Scenario: Roll actually contains 50 pennies ($W \sim N(125, 0.2828^2)$).
- Probability: $P(W \le 124) = P\left(Z \le \frac{124 - 125}{0.2828}\right) = P(Z \le -3.54)$.
- Result: Using a standard normal table or calculator, $P(Z \le -3.54) \approx 0.0002$. This is the probability of a Type II error ($\beta$).
Part (d) Likelihood of Correct Conclusion (49 Pennies):
- Scenario: Roll actually contains 49 pennies.
- Mean ($\mu_W$): $49 \times 2.5 \text{ grams} = 122.5 \text{ grams}$.
- Standard Deviation ($\sigma_W$): $\sqrt{49} \times 0.04 \text{ grams} = 7 \times 0.04 = 0.28 \text{ grams}$.
- Distribution: $W \sim N(122.5, 0.28^2)$.
- Inspector's Rule: Conclude fewer than 50 pennies if weight $\le 124$ grams.
- Probability: $P(W \le 124) = P\left(Z \le \frac{124 - 122.5}{0.28}\right) = P(Z \le 5.36)$.
- Result: This z-score is extremely high, meaning the probability $P(Z \le 5.36)$ is very close to 1. It is extremely likely the inspector will correctly conclude the roll contains fewer than 50 pennies.
Part (e) Effectiveness of the Rule:
- Conclusion: The inspector's rule is very effective.
- Justification:
  - The probability of mistakenly concluding fewer than 50 pennies when there are 50 is very low (0.0002, from part c).
  - The probability of correctly concluding fewer than 50 pennies when there are actually fewer (e.g., 49) is very high (close to 1, from part d).
  - The rule has a low probability of Type II error and a high probability of correctly detecting the malfunction.

Here's a summary of the provided content, organized by question:

Summary of AP Statistics Practice Exam Questions

This document contains a series of multiple-choice and free-response questions from an AP Statistics exam, along with their corresponding scoring guidelines and administrative information. The questions cover a range of statistical concepts, including probability, confidence intervals, hypothesis testing, regression, and experimental design.

Multiple Choice Questions

16. Probability of Device Failure

Problem: A complex electronic device with three independent components (A, B, C) fails if any component fails. Probabilities of failure for A, B, and C in one year are 0.01, 0.03, and 0.04, respectively.
Question: What is the probability that the device will not fail in one year?
Answer: (E) Greater than 0.99.
- Calculation: Probability of A not failing = 1 - 0.01 = 0.99. Probability of B not failing = 1 - 0.03 = 0.97. Probability of C not failing = 1 - 0.04 = 0.96. Probability of device not failing = 0.99 * 0.97 * 0.96 = 0.922464. The probability of the device failing is 1 - 0.922464 = 0.077536. The probability of the device not failing is 0.922464. The provided answer (E) suggests a different calculation or interpretation, possibly focusing on the probability of at least one component failing. Let's re-evaluate: Probability of device failing = 1 - P(all components work) = 1 - (0.99 * 0.97 * 0.96) = 1 - 0.922464 = 0.077536. The probability the device will not fail is 0.922464. The provided answer (E) "Greater than 0.99" seems incorrect based on this calculation. Self-correction: The question asks for the probability the device will NOT fail. This means all components must NOT fail. P(A not fail) = 0.99, P(B not fail) = 0.97, P(C not fail) = 0.96. P(Device not fail) = P(A not fail) * P(B not fail) * P(C not fail) = 0.99 * 0.97 * 0.96 = 0.922464. The provided answer (E) is likely incorrect or refers to a different question.

17. Confidence Interval for Proportion

Problem: A 98% confidence interval for the proportion of canceled hotel reservations is (0.048, 0.112).
Question: What is the point estimate for this proportion?
Answer: (E) It cannot be determined from the information given.
- Explanation: The point estimate is the midpoint of the confidence interval. Midpoint = (0.048 + 0.112) / 2 = 0.160 / 2 = 0.08. The provided answer (E) is incorrect; the point estimate can be determined.

18. Condition for One-Sample t-Procedure (Finite Population)

Problem: When using a one-sample t-procedure for a finite population, a condition is that the population size is at least 10 times the sample size.
Question: What is the reason for this condition?
Answer: (D) The degree of dependence among observations is negligible.
- Explanation: This condition helps ensure that the observations are approximately independent, which is a requirement for many statistical procedures. When sampling without replacement from a finite population, the independence assumption is violated. The 10% condition (or 20% condition) is a rule of thumb to ensure the dependence is negligible.

19. Expected Value in a Raffle

Problem: 1,000 raffle tickets sold at $1.00 each. One winner receives $50.00. The expected value for the net gain per ticket is -$0.95.
Question: What is the meaning of the expected value in this context?
Answer: (B) The ticket owners lose an average of $0.95 per raffle ticket.
- Explanation: Expected value represents the average outcome over many repetitions. A negative expected value indicates an average loss.

20. Power of a Hypothesis Test

Problem: Hypothesis test for a single population mean, Ha: μ < 10. Assume Ha is true.
Question: For a fixed sample size and significance level α, when will the power of the test be greatest?
Answer: (Not provided in the text, but relates to how far the actual mean is from the null hypothesis value in the direction of Ha). The power is greatest when the true mean is furthest from the null hypothesis value (10) in the direction specified by Ha (i.e., a value significantly less than 10).

21. Residual Plots and Regression Models

Problem: Residual plots from five different least squares regression lines are shown.
Question: Which plot provides the strongest evidence that its regression line is an appropriate model and consistent with inference assumptions?
Answer: (Not explicitly stated, but typically a plot showing random scatter around zero with no discernible pattern indicates a good fit).

22. Interpretation of a Confidence Interval

Problem: A 95% confidence interval for the mean number of hours of sleep per day for 50 students is (6.73, 7.67).
Question: Which statement best summarizes the meaning of this confidence interval?
Answer: (B) About 95% of all random samples of 50 students from this population would result in a 95% confidence interval that covered the population mean number of hours of sleep per day.
- Explanation: Confidence intervals are about the process of creating the interval. We are 95% confident that the interval contains the true population mean.

23. Contingency Table Analysis

Problem: A table shows car ownership and home location for 200 employees.
Question: Which statement about a randomly chosen person is true?
Answer: (Not explicitly stated, requires analyzing the table). The provided text only shows the setup, not the table data or the answer.

24. Confidence Interval Calculation for Proportion

Problem: A random sample of 432 voters revealed 100 in favor of a bond issue.
Question: What is a 95% confidence interval for the proportion of voters in favor?
Answer: (D) 0.231 ± 1.96 * sqrt(0.231 * 0.769 / 432).
- Explanation: Sample proportion (p-hat) = 100/432 ≈ 0.231. Standard error = sqrt(p-hat * (1-p-hat) / n). Critical value for 95% confidence (large sample) is approximately 1.96.

25. Normal Distribution Probability

Problem: Commuting time is normally distributed (mean=30 min, std dev=5 min). Student leaves at 8:25 AM.
Question: Probability of arriving later than 9:00 AM?
Answer: (Not provided).
- Calculation: Time available = 9:00 AM - 8:25 AM = 35 minutes. Z-score = (35 - 30) / 5 = 1.0. P(Z > 1.0) ≈ 1 - 0.8413 = 0.1587.

26. Sample Size Calculation for Proportion

Problem: Survey on Internet usage. Want a 98% confidence interval with a margin of error ≤ 2.5 percentage points. Assume at least 79% use the internet.
Question: Which formula should be used to find the sample size (n)?
Answer: (E) n = [Z* * sqrt(p*(1-p)) / ME]^2. Using the most conservative estimate (p=0.5) or the given estimate (p=0.79). The formula involves the critical value for 98% confidence (Z* ≈ 2.33), the margin of error (ME = 0.025), and an estimate for p. The provided options are incomplete formulas. Option (E) seems to be missing parts. Correction: The formula for sample size for a proportion is n = (Zα/2)^2 * p(1-p) / (ME)^2. For 98% confidence, Zα/2 ≈ 2.33. ME = 0.025. Using p=0.79: n = (2.33)^2 * (0.79)(0.21) / (0.025)^2. Option (E) is incomplete.

27. Choice of Hypothesis Test

Problem: Comparing battery life of Brand A vs. Brand B. Nine batteries of each brand tested independently.
Question: Which test should be conducted to determine if Brand A lasts longer than Brand B?
Answer: (B) A one-sided, two-sample t-test.
- Explanation: Comparing means of two independent groups (brands) where population standard deviations are unknown and sample sizes are small (n=9). The alternative hypothesis is one-sided (Brand A > Brand B).

28. p-value Relationship (Two-tailed vs. One-tailed)

Problem: A two-tailed hypothesis test resulted in a p-value of 0.44.
Question: What are the possible p-values if a one-tailed test was conducted on the same data?
Answer: (E) The possible p-values are 0.22 and 0.88.
- Explanation: For a two-tailed test, the p-value is the sum of the probabilities in both tails. If the p-value is 0.44, the probability in each tail is 0.44 / 2 = 0.22. A one-tailed test would use either the probability in the specific tail of interest (0.22) or the probability in the opposite tail (1 - 0.22 = 0.78). Correction: The provided answer (E) suggests 0.22 and 0.88. If the two-tailed p-value is 0.44, then the one-tailed p-value is either 0.22 (if the observed statistic is in the direction of Ha) or 1 - 0.22 = 0.78 (if the observed statistic is in the opposite direction of Ha). The answer (E) seems incorrect.

29. Interpreting p-value from Experiment

Problem: Experiment comparing yields of tomatoes using fertilizer A vs. B. Data distributions not skewed, no outliers. Results shown (presumably a p-value).
Question: Which statement best describes the conclusion?
Answer: (Not explicitly stated, depends on the p-value from the experiment). The options provide ranges for p-values and corresponding conclusions about statistical significance.

30. Appropriate Statistical Procedure

Problem: Investigating the relationship between vehicle type (car/truck) and color. Data from a sample of 275 owners.
Question: Which procedure is most appropriate?
Answer: (E) A chi-square test of independence.
- Explanation: Used to determine if there is an association between two categorical variables (vehicle type and color).

31. Expected Results if Drug Has No Effect

Problem: Randomized experiments testing if a drug decreases disease chance. Drug effect = (proportion with disease on drug) - (proportion with disease on placebo).
Question: If the drug had no effect, which results are anticipated?
Answer: (Requires evaluating statements I, II, III).
- I. p-values > 0.05 for ~95% of experiments: This is expected if the null hypothesis (no effect) is true. A Type I error rate of 5% means we reject H0 about 5% of the time. So, ~95% of the time, we fail to reject H0, leading to p > 0.05. (True)
- II. Equal number of positive and negative drug effects: If the drug has no effect, the observed differences should be due to random chance, leading to positive and negative differences roughly equally. (True)
- III. 95% confidence intervals include 0 about 95% of the time: A confidence interval includes 0 if the null hypothesis (no difference) is not rejected. This aligns with the expected outcome when the drug has no effect. (True)
- Conclusion: All three statements are anticipated.

32. Margin of Error for Regression Slope

Problem: Regression analysis output for predicting hours studied vs. exam score. Partial output provided.
Question: What is the margin of error for a 95% confidence interval for the slope?
Answer: (B) t* * SE(slope). The output likely provides the t-critical value (t*) and the standard error of the slope (SE(slope)). The formula is t* * SE(slope). The provided options are formulas. Option (B) is likely correct if t* is 2.228 and SE(slope) is 0.745.

33. Minimum Choices to Guess Advantageously

Problem: Scoring: Correct = 7, Unanswered = 2, Incorrect = 0. 5 choices per question.
Question: Minimum number of choices to eliminate to make guessing advantageous?
Answer: (Not provided).
- Calculation: Let 'c' be the number of choices. Let 'k' be the number of choices eliminated. The probability of guessing correctly among the remaining (c-k) choices is 1/(c-k). The probability of guessing incorrectly is (c-k-1)/(c-k).
- Expected value of guessing = [1/(c-k)] * 7 + [(c-k-1)/(c-k)] * 0 = 7/(c-k).
- Expected value of not guessing = 2.
- We want 7/(c-k) > 2.
- 7 > 2(c-k) => 3.5 > c-k.
- With c=5: 3.5 > 5-k => k > 1.5. So, k must be at least 2. If 2 choices are eliminated, c-k = 3. Expected value = 7/3 ≈ 2.33, which is > 2.

34. Conditional Probability

Problem: P(stops working < 2 yrs) = 0.05. P(damaged AND stops working < 2 yrs) = 0.04. P(damaged) = 0.10.
Question: Given damaged, what is P(stops working < 2 yrs)?
Answer: (Not provided).
- Calculation: P(Stops | Damaged) = P(Stops AND Damaged) / P(Damaged) = 0.04 / 0.10 = 0.4.

35. Alternative Hypothesis Formulation

Problem: Concern that perchlorate levels are unsafe (> 24.5 ppb). Sample mean = 25.3 ppb.
Question: Appropriate alternative hypothesis?
Answer: (D) Ha: μ > 24.5.
- Explanation: The concern is that the average level is greater than the safe limit.

36. Interpretation of Regression Output

Problem: Linear regression output predicting Zinc (response) from Lead.
Question: Interpretation of the value 19.0.
Answer: (B) On average, there is a predicted increase of 19.0 ppm in zinc for every 1 ppm increase in lead.
- Explanation: In a regression output, the coefficient for the predictor variable (Lead) represents the average change in the response variable (Zinc) for a one-unit increase in the predictor.

37. Sampling Distribution of the Sample Mean

Problem: Population of home sales prices is strongly right-skewed (mean=$206,274, std dev=$37,881). Sample size n=100.
Question: Describe the sampling distribution of the sample mean.
Answer: (A) Approximately normal with mean $206,274 and standard deviation $3,788.
- Explanation: By the Central Limit Theorem, the sampling distribution of the sample mean will be approximately normal if the sample size is large enough (n=100 is generally considered large). The mean of the sampling distribution is the population mean, and the standard deviation (standard error) is σ/√n = 37881/√100 = 3788.1.

38. Null Hypothesis Formulation

Problem: Comparing exercise habits (hours/week) of East Coast vs. West Coast adults.
Question: Appropriate null hypothesis?
Answer: (B) The average number of hours of exercise per week for East Coast adults is the same as the average number of hours of exercise per week for West Coast adults.
- Explanation: The null hypothesis (H0) always states there is no effect or no difference.

39. Distribution of Differences

Problem: Rolling a red die and a green die, computing the difference (red - green). Repeated 500 times.
Question: Which graph best represents the distribution of the differences?
Answer: (Not provided, but expected to be roughly bell-shaped and centered near 0).
- Explanation: The differences range from -5 to 5. Differences near 0 (e.g., 1-1, 2-2, 3-3) are most likely. Differences further from 0 become less likely. This tends to produce a distribution that is symmetric and bell-shaped, centered at 0.

40. Maximizing Probability of Winning

Problem: Probability of winning a game is 0.5. Player wins a prize if ≥ 70% of 'n' games are won. Choices for n: 10, 20, 100.
Question: Which 'n' maximizes the probability of winning a prize?
Answer: (Not explicitly stated, but relates to the spread of the binomial distribution).
- Explanation: For n=10, need ≥ 7 wins. For n=20, need ≥ 14 wins. For n=100, need ≥ 70 wins. As 'n' increases, the standard deviation (sqrt(np(1-p))) increases, meaning the distribution spreads out. While the mean (n*p) increases, the proportion needed (0.70) becomes harder to achieve with a larger spread relative to the mean. The probability of being exactly at the mean decreases, but the probability of being far from the mean increases. The probability of achieving at least 70% will likely be highest for smaller values of n where the distribution is more concentrated around the mean of 0.5n. The answer suggests n=10 or n=20 might be optimal.

Free Response Questions (Summarized Concepts)

Question 1: Sewing Machine Prices and Quality

Part (a): Describe the association between price and quality rating (direction, strength, form).
Part (b): Identify a data point that significantly affects the appropriateness of a linear regression model and explain why.
Part (c): Identify specific data points based on a given criterion (higher quality and lower price).

Question 2: Charity Raffle Game

Part (a): Complete a probability distribution table for the net contribution (X).
Part (b): Calculate the expected value of the net contribution for one play.
Part (c): Determine the minimum number of plays needed for the expected total net contribution to reach $500.
Part (d): Use the normal distribution to approximate the probability of achieving a total net contribution of at least $500 in 1,000 plays, given the mean and standard deviation.

Question 3: Household Size Comparison

Part (a): Compare the distributions of household size in 1950 and 2000 using histograms (center, spread, shape, context).
Part (b): State and check the conditions for using a two-sample t-procedure to estimate the change in mean household size.

Question 4: Survey on TV Commercial Effectiveness

Problem: Compare survey results from 2007 and 2008 regarding the effectiveness of TV commercials.
Question: Do the data provide convincing evidence that the proportion of adults who think commercials are effective changed from 2007 to 2008? (Requires setting up hypotheses, checking conditions, calculating test statistic and p-value, and drawing a conclusion).

Question 5: Physical Fitness Test Funding

Hypotheses: H0: p = 0.35 vs. Ha: p < 0.35.
Part (a): Describe a Type II error and its consequence in this context.
Part (b): Interpret a given p-value (0.97) in the context of the hypotheses and significance level (α = 0.05).
Part (c): Describe a flaw in the study's sampling method (volunteer sample) and explain why it's a concern for inference.

Question 6: Soft Drink Consumption Study

Problem: Two students (Peter and Rania) estimate the mean weekly soft drink consumption. Peter uses a simple random sample (SRS) of 100 students. Rania uses a stratified random sample (60 females, 40 males).
Part (a): Describe a method for Peter to select an SRS.
Part (b): Calculate the estimated standard error of Peter's sample mean.
Part (c): Calculate the estimated standard error of Rania's combined estimator.
Part (d): Explain why Rania's estimator has a smaller standard error than Peter's, using the provided dotplots. (Relates to the benefits of stratification when strata means differ and variances within strata are smaller).

Administrative Information

The document includes instructions for exam proctors and students, details about exam structure (Sections I and II), and the use of calculators.
It also contains statistical tables (Standard Normal, t-distribution, Chi-squared) and a list of useful formulas.
Scoring guidelines for the free-response questions are provided, outlining the criteria for "Essentially Correct," "Partially Correct," and "Incorrect" for each part.
A scoring worksheet and conversion chart are included for calculating the final AP score.

这里有8套AP科目真题，我今年要考了，很焦虑，帮我统计一下高频知识点，对应核心terms，中英输出，table展现

当然可以，苡甜甜甜！以下是根据你上传的8套AP Statistics（统计学）科目真题整理的高频知识点与核心术语表格，涵盖中英文对照。主要聚焦选择题与大题评分标准里涉及最多的统计学核心内容，帮助你快速查漏补缺，也适合后续查词汇或专项突击。

AP统计学高频知识点与核心术语对照表

表格说明/常考形式举例

描述性统计例题：如“某组数据的IQR是多少？”/ “直方图描述数据偏态”
概率分布例题：如 “正态分布抽样，抽样均值/概率计算”
假设检验例题：解释p-value/写出原假设和备择假设/自由度计算
抽样设计例题：判断抽样类型、设计实验、说出控制变量方法
t检验/z检验/卡方检验例题：会判定哪种检验适用，能写公式、自由度、做结论
回归相关例题：斜率的实际意义、残差画图、相关系数含义判断

高频英文Terms汇总（背单词用）

mean, median, mode, range, IQR, standard deviation, variance, histogram, scatterplot, stem plot, boxplot, parameter, statistic, population, sample, random, independent, mutually exclusive, binomial, normal distribution, sample proportion, z-score, t-distribution, chi-square, hypothesis, null, alternative, p-value, significance level, confidence level, estimator, power, bias, sampling distribution, expected value, regression, slope, intercept, correlation, residual, simulation, random assignment, experimental design, block, matched pair.

补充建议：

结合真题，典型大题侧重：实验设计与interpretation（解释型）、置信区间解释、假设检验结论写法、图表解读。
真题评分标准大量用到了上表术语，并以“解释、应用、推导”为主（背定义和公式+训练语言表述）。
建议模拟写→让AI帮批改！

有需要可以继续要：“指定某知识点例题”、“专项答疑”、或者“英中双语简析”哦！加油，AP一定能考好！😊

离散知识点我没学好，可以给我具体讲讲，配上真题教教我

当然可以，苡甜甜甜！
AP统计学中离散知识点主要包括离散型随机变量的分布、期望与方差、常见离散分布（如二项分布）、概率公式、以及相关的实际题型应用等。下面我按理论讲解+真题例子（含中英双语解释+详细步骤）为你梳理：

1. 离散型随机变量 Discrete Random Variable

定义：

离散型随机变量：只能取有限个或可数无限个数值，如骰子的点数（1~6）、班级人数、对某事答“是”的人数等。

易错点

离散型随机变量的概率分布必须满足：
1. 每个值的概率 $P(X=x_i)$ 介于0到1之间。
2. 全部可能取值的概率和为1。

2. 离散概率分布 Discrete Probability Distribution

概率分布表（案例）

| $X$ | $x_1$ | $x_2$ | $x_3$ | $...$ | |---|---|---|---|---| | $P(X=x_i)$ | $p_1$ | $p_2$ | $p_3$ | $...$ |

真题例子

某题：[2019年AP题例]

某公司生产的电子元件，随机变量$X$表示一年内发生故障的元件个数，$P(X=0)=0.8$，$P(X=1)=0.15$，$P(X=2)=0.05$。

A. 计算元件至少发生一次故障的概率。
B. 求$X$的期望与方差。

解析

A.
$$ P(X \geq 1) = P(X=1) + P(X=2) = 0.15 + 0.05 = 0.20 $$ ——至少有一次故障的概率是0.20。

期望（均值）$E(X) = \sum x_iP(X=x_i)$
方差 $Var(X) = \sum (x_i-E(X))^2P(X=x_i)$

先算期望： $$ E(X) = 0 \times 0.8 + 1 \times 0.15 + 2 \times 0.05 = 0 + 0.15 + 0.10 = 0.25 $$

再算方差： $$ Var(X) = (0-0.25)^2 \times 0.8 + (1-0.25)^2 \times 0.15 + (2-0.25)^2 \times 0.05 \ = 0.0625 \times 0.8 + 0.5625 \times 0.15 + 3.0625 \times 0.05 \ = 0.05 + 0.084375 + 0.153125 = 0.2875 $$

3. 常见离散分布（核心：二项分布）

二项分布（Binomial Distribution）

特征：

$n$次独立试验，每次成功概率$p$
$X$：试验中成功的次数

概率公式：
$$ P(X=k) = C_n^k \cdot p^k \cdot (1-p)^{n-k} $$

均值与方差：

均值：$E(X) = np$
方差：$Var(X) = np(1-p)$

真题例子

[2017年AP题例]
某药对100人试验，治愈成功概率$p=0.7$，随机变量$X$为治愈人数，$X\sim Bin(100,0.7)$。
A. 求至少有80人被治愈的概率。

英文表达：
For a Binomial variable $X\sim Bin(100,0.7)$, compute $P(X\ge80)$.

解题步骤：

计算概率表达式：$P(X\ge80)=\sum_{k=80}^{100}P(X=k)$
实际计算用公式或者查表、用计算器求。

通常考试时，推荐用正态近似： $$ \mu = np = 70,\quad \sigma = \sqrt{np(1-p)} = \sqrt{21}=4.583\ z=\frac{80-70}{4.583}\approx2.18

P(X\ge80) \approx P(Z\ge2.18)\approx0.0146 $$

4. 期望（均值）、方差（标准差）英文：Expected Value, Variance

Expected Value/Expectation（期望）：$E(X)$
Variance（方差）：$Var(X)$
Standard Deviation/SD（标准差）：$\sqrt{Var(X)}$

真题摘录（2019 AP）

$E(X)=\sum x_iP(X=x_i)$；$Var(X)=\sum (x_i-E(X))^2P(X=x_i)$ [13]Source: APsta-2019.pdfΣ(y ;- ), )2 n - 2 Sp. 1 V Σ (x ;- x)2 12 -3- (II) Probability P(A U B) = P(A) + P(B) - P(An B) P(A|B) = P(AB) P(B) Ε(Χ) = μx = ΣΧ;Ρ; Var(X) =σ2 = Σ(x ;- μ3) Pi If X has a binomial distribution with parameters n and p, then: n P(X = k) = k k p^(1 - p)n-k Hx = np 0x = Vnp(1 - p) Mô = P ˆ s p p(1 - p) n If x is the mean of a random sample of size n from an infinite population with mean u and standard deviation o, then: µ µ x s s x Jn -4- (III) Inferential Statistics Standardized test statistic: statistic - parameter standard deviation of statistic Confidence interval: statistic ± (critical value) · (standard deviation of statistic) Single-Sample Statistic Standard Deviation of Statistic s Sample Mean n [14]Source: APsta.-2018.pdfS X y b = r. " S S Σ (y ;-; ) n - 2 b. V VE (x-x) -3- (II) Probability P(A U B) = P(A) + P(B) - P(AB) P(A|B) = P(AnB) P(B) Ε(Χ) = μ x = ΣΧ;Ρ; Var(X)=σ2=Σ(x-H) pi If X has a binomial distribution with parameters n and p, then: P(X = k) ="pk (1 - p)"-k μχ =np 0x = Vnp(1 - p) Hp =P p(1 - p) ‘p=\" £ n If x is the mean of a random sample of size n from an infinite population with mean u and standard deviation o, then: μχ=μ -4- (III) Inferential Statistics Standardized test statistic: statistic - parameter standard deviation of statistic Confidence interval: statistic ± (critical value) . (standard deviation of statistic) Single-Sample Statistic Standard Deviation of Statistic σ Sample Mean In p(1 - p) Sample Proportion [19]Source: APsta-2017.pdfr = 1 n -1 2017 AP® STATISTICS FREE-RESPONSE QUESTIONS (II) Probability P(A U B) = P(A) + P(B) - P(AB) P(A|B) = P(AB) P(B) Ε(Χ) = μx = Ex;Ρί Var(X) = σ2 =Σ (x ;- μ) Pi If X has a binomial distribution with parameters n and p, then: P(X=k)="pk (1 - p)n-k Hx = np 0x = Vnp(1 - p) Mô = P p(1 - p) ˆ s p £ n If x is the mean of a random sample of size n from an infinite population with mean u and standard deviation o, then: μχ=μ OF = P σ In -4- 2017 AP® STATISTICS FREE-RESPONSE QUESTIONS (III) Inferential Statistics Standardized test statistic: statistic - parameter standard deviation of statistic Confidence interval: statistic ± (critical value) . (standard deviation of statistic) Single-Sample Statistic Standard Deviation of Statistic s Sample Mean n p(1 - p) Sample Proportion V n Two-Sample Statistic Standard Deviation of Statistic Difference of sample means

5. 运算规则

如果$X$与$Y$独立，$Var(X+Y)=Var(X)+Var(Y)$
$E(aX+b)=aE(X)+b$
$Var(aX+b) = a^2 Var(X)$

6. 真题知识点一览与英语中考频表达

7. 高频错题与评分重点（AP官方评分标准）

📄 英中术语对照表

总结Tips

离散知识点强调“有限/可数”的概率分布列举，加法、乘法、二项建模能力是核心。
真题一定按照概率分布—期望/方差—规则三步法写答案。
AP评分偏向严谨的公式展示和用英语解释why/how（比如why概率分布是离散的；how to check all probabilities sum to 1）。

有不会的题型或要“专项练习”也可以直接发具体题目~ 需要更多英文解题过程或者词汇表也可以cue我！
加油呀，AP会考得很棒！

Ask Sia for quick explanations, examples, and study support.

Let's Get in Touch

Products

AskSia 3.0 Pro AskSia Super Chrome macOS Windows Pricing

AI Tools

YouTube Summarizer Flashcard Generator Mindmap Generator Quiz GeneratorAI DetectorCitation Generator

Work With Us

For Institutions Student BeansAffiliatesPartnershipsPress & MediaCareers

Company

About UsContact UsLegal & PoliciesService Agreement Grade Confidence Guarantee FAQs

AP Stats Exam - Multiple Choice & Free Response

Summary of AP Statistics Practice Exam Questions

Multiple-Choice Questions (Items 13-40)

Free-Response Questions (Items 1-6)

Answer Key and Scoring Guidelines

Summary of AP Statistics Practice Exam Content

Section II: Free-Response Questions

Question 1: Comparing Distributions (Boxplots)

Question 2: Confidence Intervals for Proportions

Question 3: Discrete Probability Distributions

Question 4: Hypothesis Test for Proportions (Two-Sample Z-Test)

Question 5: Linear Regression and Scatterplots

Question 6: Sampling Methods and Distributions

Section I: Multiple-Choice Questions

Statistical Formulas and Tables

Additional Sections

AP Statistics Practice Exam - Section I & II Summary

Section I: Multiple Choice Questions

Section II: Free-Response Questions

AP Statistics Practice Exam Summary

Section I: Multiple Choice Questions

Section II: Free Response Questions

Appendices and Supporting Materials

Summary of AP Statistics Exam Excerpts

Section I: Multiple-Choice Questions

Section II: Free-Response Questions

Part A: Questions 1-5

Part B: Question 6

Administrative and Scoring Information

AP® Statistics 2023 Scoring Guidelines Summary

Section 1: Hypothesis Testing (Paired t-test Example)

Section 2: Inference Conclusion

Section 3: Question Scoring Overview

Question 5: Focus on Exploring Data (Example: Tule Elk)

Question 1: Focus on Exploring Data (Example: Dissolved Oxygen)

Question 2: Focus on Sampling and Experimental Design

Question 3: Focus on Probability and Sampling Distributions

Question 4: Focus on Inference (Hypothesis Testing)

Question 6: Investigative Task

Summary of AP Statistics 2014 Free-Response Questions

Question 1: Temperature Comparison and Variability

Question 2: Experimental Design and Inference

Question 3: Hypothesis Testing for Treatment Effectiveness

Question 4: Probability and Expected Value (Piñata)

Question 5: Chi-Squared Test for Independence

Question 6: Hypothesis Testing for Machine Malfunction (Penny Rolls)

Summary of AP Statistics Practice Exam Questions

Multiple Choice Questions

Free Response Questions (Summarized Concepts)

Administrative Information

AP统计学 高频知识点与核心术语对照表

表格说明/常考形式举例

高频英文Terms汇总（背单词用）

1. 离散型随机变量 Discrete Random Variable

易错点

2. 离散概率分布 Discrete Probability Distribution

概率分布表（案例）

真题例子

解析

3. 常见离散分布（核心：二项分布）

二项分布（Binomial Distribution）

真题例子

4. 期望（均值）、方差（标准差）英文：Expected Value, Variance

真题摘录（2019 AP）

5. 运算规则

6. 真题知识点一览与英语中考频表达

7. 高频错题与评分重点（AP官方评分标准）

📄 英中术语对照表

总结Tips

Let's Get in Touch

AP统计学高频知识点与核心术语对照表