Lecture Notes: Understanding Hypothesis Testing and Confidence

Introduction

• Discuss alternate scenarios and the main idea of hypothesis testing.
• Use of z-scores for probability calculations.

Scenarios and Calculations

• Initial Example:

  • If z-score = 2.28, but consider if z-score = 1.7.
  • Mu (mean) = 4 is the hypothesis.
  • Compare to new x-bar value of 4.2.
  • Initially 98% confident that mu > 4.

• New Sample Observations:

  • With x-bar = 4.2, confidence in mu = 4 is adjusted.
  • New z-score calculation (z = 0.9128) leads to 81.93% confidence.
  • Confidence that mu > 4 decreases.

• Shading and Confidence:

  • Unshaded area = confidence.
  • Shaded area represents the opposite scenario.

Further Example with Higher x-bar

• Example with x-bar = 4.8:
  • Confidence increases to 99.99% that mu > 4.
  • Further from original guess means less confidence in mu = 4.
  • Differences in data from the hypothesis alter confidence.

Discussion on Hypothesis Testing

• Hypothesis Test Process:

  • Make a claim about the population (mu).
  • Collect data (x-bar).

• Confidence vs. P-value:

  • Confidence = unshaded area (e.g., "fat middle").
  • P-value = shaded area.
  • Relationship: confidence + P-value = 1.

• Terminology:

  • P-value introduced as a way to interpret confidence.

Rejecting or Accepting Hypotheses

• Interpretation of Confidence:

  • Example: 98.88% confidence that mu > 4.
  • Reject null hypothesis (H0) that mu = 4 in favor of alternative hypothesis (H1) that mu > 4.

• Traditional vs. Modern Approach:

  • Reject/accept method traditional, yet declining in favor.
  • Confidence approach preferred for clarity.

Conclusion

• Understanding hypothesis testing involves grasping confidence, P-value, and interpreting results.
• Different ways to express findings: confidence, P-value, and accept/reject method.