LWB Level 3 Inference 3.10 Learning Workbook

The LWB Level 3 Inference 3.10 Learning Workbook is a specialized, write-on resource designed for the NCEA Level 3 Internal Assessment: Investigate a given multivariate data set. This workbook guides students through the "Statistical Enquiry Cycle," focusing on moving beyond simple observations to making formal, justified conclusions about a wider population using modern re-sampling methods.

Key Features

• The PPDAC Cycle for Inference: A structured approach to the internal assessment task, ensuring students meet all requirements for the "Problem, Plan, Data, Analysis, and Conclusion" stages.

• Problem Formulation: Learning how to craft a precise, population-based comparative question (e.g., "I wonder if the median birth weight of babies born to non-smoking mothers is greater than the median birth weight of babies born to smoking mothers in New Zealand...").

• Data Cleaning and Exploration:

  • Sampling Theory: Understanding the importance of representative samples and identifying potential biases in the data.

  • Variables: Distinguishing between categorical and numerical variables to ensure the correct analysis is performed.

• Visualizing Data with Box Plots: Mastery of the Box and Whisker plot. Students learn to interpret the "middle 50%" (Interquartile Range), look for overlaps, and identify unusual values or skewness in the distribution.

Advanced Statistical Reasoning

• Bootstrapping (The Confidence Interval): The core of Level 3 Inference. Instead of relying on traditional formulas, students use computer-assisted re-sampling (typically via NZGrapher) to create a "Bootstrap Confidence Interval."

  • The Process: Learning how the software "re-samples with replacement" thousands of times to estimate the variability of the median difference.

  • The "Rule of Thumb": Interpreting the resulting interval to determine if we can make a call about the population. If zero is not in the interval, a difference likely exists.

• Sampling Variability: Understanding that a different sample would result in a different estimate, and how the bootstrap interval accounts for this uncertainty.

• Population Generalization: Learning to carefully define the "target population" and discussing whether the sample data is sufficient to make a broad claim.

Workbook Highlights

• NZGrapher Integration: Features step-by-step screenshots and guides for generating Bootstrap distributions and high-quality statistical graphs.

• Internal Assessment Template: Provides a clear report-writing structure, helping students organize their statistical evidence into a format that meets the criteria for Excellence.

• Contextual Data Sets: Includes authentic New Zealand data sets—ranging from environmental metrics to social statistics—to practice analysis on real-world scenarios.

• Critiquing and Reflecting: Exercises designed to help students discuss the limitations of their investigation and the impact of sample size on the precision of their confidence intervals.

• Full Answer Appendix: Provides model answers for the analysis and conclusion sections, showing students how to write with the required technical precision and depth.

• Glossary of Inference Terms: A complete guide to essential vocabulary—such as Point Estimate, Confidence Interval, Sampling Error, and Re-sampling—to ensure students use professional statistical language.