AP Stat FRQ Practice

Section II: Free-Response

1. Khaled’s Reading Assignment

  • Context: Khaled is tracking reading time and pages left among classmates.
  • Data Collection: Each student records hours reading and pages remaining.
  • Scatterplot Description:
      - Assessment of Relationship: Examine correlation between hours read and pages left.
  • Regression Analysis:
      - Regression Equation: extnumberofpagesleft=592.50421.512imes(exthoursspentreading)ext{number of pages left} = 592.504 - 21.512 imes ( ext{hours spent reading})
      - Slope Interpretation:
        - The slope 21.512-21.512 indicates that for each additional hour spent reading, the number of pages left decreases by approximately 21.5 pages.
  • Khaled’s Reading Performance:
      - Khaled read for 4 hours and has 540 pages left.
      - (i) Residual Calculation:
        - Predicted pages left: 592.504(21.512imes4)592.504 - (21.512 imes 4)
        - Residual: 540extpredictedpagesleft540 - ext{predicted pages left}, calculated value to be shown.
      - (ii) Residual Interpretation:
        - The residual indicates how much Khaled’s reading performance deviated from the prediction made by the model. A negative residual suggests he has fewer pages left than predicted for his reading time, indicating efficient reading.

2. Psychologist Experiment on Lighting Conditions

  • Research Objective: Investigate lighting effects on work productivity.
  • Methodology:
      - Sample Size: 40 employees randomly selected.
      - Assignments: 20 in bright light, 20 in dim light.
      - Response Measure: Quality rated from 0 (low) to 5 (high).
  • Study Components:
      - Treatments: Bright light vs. dim light.
      - Experimental Units: 40 employees in total.
      - Response Variable: Quality of work, rated from 0 to 5.
  • Generalizability of Results:
      - Threatened by the small and specific sample size; may not represent all employees in the corporation due to unknown variance in productivity across different roles or departments.
  • Proposed Change in Design:
      - Design Type: Matched pairs design.
      - Statistical Advantage: Reduces variability by pairing similar employees, thus isolating the effect of lighting conditions more clearly.

3. Airline Flight Delay Probability Analysis

  • Claimed Distribution: Minutes late (mean = 2.4 minutes, SD = 2.1 minutes).
  • Delay Definition: More than 5 minutes late indicates a delay.
  • (i) Probability Calculation:
      - Find P(X > 5) using the normal distribution with the stated parameters:
        - Standardizing: Z=XextmeanextSDZ = \frac{X - ext{mean}}{ ext{SD}}
        - Specifically: Z=52.42.1Z = \frac{5 - 2.4}{2.1}
  • Kelly’s Flights Monitoring:
      - (ii) Random Variable Definition: Let YY denote the number of delayed flights, where YextfollowsB(20,p)Y ext{ follows } B(20, p), with p = P(X > 5).
      - Expectation of Delayed Flights: Expect 20p20p delayed flights based on the value calculated from the probability.
  • Kelly’s Experience of Delays:
      - (i) Probability of 7 or More Delayed Flights:
        - Use binomial probability formula or normal approximation:
        - Calculate P(Yext7)P(Y ext{ ≥ } 7) using binomial (or normal approx. if applicable).
      - (ii) Evidence Review:
        - Compare Kelly's observations against the expected probability from the airline claims to assess if her experience provides statistical evidence of more frequent delays.

4. Educational Research on Teaching Methods

  • Research Design: 100 students paired by previous test scores, assigned to two different teaching methods.
  • Method Details:
      - Pairing Basis: Two lowest scores paired, the two highest scores paired, etc.
      - Method Comparison: New vs. traditional.
  • Findings Summary:
      - Sample Size: Both methods have n = 50.
      - Mean Scores: New (mean = 81.4), Traditional (mean = 78.1), Difference (3.3).
      - Standard Deviations: New (8.6), Traditional (6.5), Difference (7.7).
  • Statistical Evidence Evaluation:
      - Analyze means and variances, apply significance level of extα=0.05ext{α} = 0.05 to conclude if the new method outperforms the traditional method statistically validly.

5. Attitudes towards AI Technology by Age Group

  • Survey Overview: Two age groups (18 to 44 years, 45 years and older) surveyed on AI attitudes.
  • Mosaic Plot Construction: Visual representation of responses across age groups (positive, neutral, negative).
  • Association Description:
      - Examine if attitudes significantly differ between the age groups based on summarized survey data.
  • Chi-Square Test for Homogeneity:
      - Hypotheses: Null: no difference; Alternative: difference exists.
      - Test statistic calculation: Show formula and calculate for the respective counts.
      - Interpretation of Test Statistic: Contextualize what the test statistic indicates regarding age differences in AI attitudes.
      - Observed Frequencies:
        - Positive responses: 24 (18-44) and 6 (45+).
        - Neutral: 18 and 12.
        - Negative: 6 and 14.

6. Mateo's AP Course Enrollment Study

  • Student Sampling: Mateo sampled 200 students, finding 82 enrolled in AP courses.
  • Confidence Interval Construction: Seek to estimate population proportion of students enrolled.
  • Inference Conditions Check: Confirm random sampling and success/failure condition status for validity of inferences.
  • Normal Approximation Justification:
      - Based on sample size and success/failure counts for standard normal approximation of the sampling distribution.
  • Type I Error Definition: Incorrectly rejecting the null hypothesis that the true proportion of AP enrollment equals 35%.
  • Confidence Level Impact: Higher confidence level (95% vs. 90%) results in wider intervals and lower Type I error rate, thus balanced decision making.
  • Confidence Interval Outcomes:
      - 90% confidence interval: (0.353, 0.467) - decision based on the hypothesis.
      - 95% confidence interval: (0.342, 0.478) - different conclusion may emerge based on levels.
  • P-value Range Estimate:
      - Discuss based on confidence intervals without numeric calculations, influencing statistical significance assessments.