Hypothesis Testing Overview

Chapter 6: Hypothesis Testing Notes

Objectives

  • Develop Null and Alternative Hypothesis
  • Understand Type I and Type II Errors
  • Analyze Population Means: Known and Unknown
  • Assess Population Proportion

Hypothesis Testing

  • Hypothesis testing is a method to decide whether to reject a statement about a population parameter.
  • Null Hypothesis (H₀): A tentative assumption about the population parameter.
  • Alternative Hypothesis (H₁): The opposite of the null hypothesis, representing what the researcher aims to support.
  • The process utilizes data from a sample to evaluate these competing hypotheses.

Developing Null and Alternative Hypotheses

  • The formulation of hypotheses can be challenging and context-dependent.
  • It may be easier to identify the alternative hypothesis first based on the researcher’s belief.
  • The conclusion that the alternative hypothesis is true is based on sufficient evidence against the null hypothesis.
  • Example 1: A new teaching method is believed to be better.
  • H₁: The new method is better.
  • H₀: The new method is not better.
  • Example 2: A new sales bonus plan may increase sales.
  • H₁: The new plan increases sales.
  • H₀: The new plan does not increase sales.

Understanding Errors in Hypothesis Testing

Type I Error
  • Occurs when a true null hypothesis is incorrectly rejected.
  • The probability of a Type I error is represented by the level of significance (α).
  • These types of tests are often referred to as significance tests.
Type II Error
  • Happens when a false null hypothesis is not rejected.
  • It is generally difficult to control Type II errors. Statisticians use language like "do not reject" to mitigate confusion.

Forms of Null and Alternative Hypotheses for Population Means

  • The equality (e.g., =, ≤) typically appears in the null hypothesis.
  • Common forms include:
  • H₀: μ ≤ μ₀ (Null hypothesis states the mean is less than or equal to a hypothesized value)
  • H₁: μ > μ₀ (Alternative hypothesis states the mean is greater)
  • Similar forms exist for two-tailed tests (e.g., H₀: μ = μ₀ vs. H₁: μ ≠ μ₀)

Example: Metro EMS

  • Goal: Assess if the mean response time meets the target of 12 min.
  • H₀: μ ≤ 12
  • H₁: μ > 12
  • This implies that if H₀ is accepted, no follow-up action is needed; if rejected, corrective measures may be taken.

Hypothesis Testing Steps

General Steps
  1. Develop the null and alternative hypotheses.
  2. Specify the level of significance (α).
  3. Collect sample data and compute test statistics.
  4. Calculate the p-value or critical values.
  5. Make a decision to reject or not reject the null hypothesis based on the p-value or critical value.

One-Tailed vs. Two-Tailed Tests

  • One-Tailed: Tests if a parameter is greater than or less than a specific value.
  • Two-Tailed: Tests if a parameter is significantly different (either greater or less) than a value.
  • Rejection rules depend on the computed test statistic and whether it falls into the rejection region defined by critical values or p-values.

Case Studies

Example: Hilltop Coffee
  • Claim: Test if a can contains at least 3 lbs of coffee.
  • H₀: μ ≥ 3
  • H₁: μ < 3
  • Conducted test with a sample showing a mean of 2.92 lbs; implies the company is underfilling their cans if rejected.
Example: MaxFlight Inc.
  • Claim: Test if the mean driving distance of golf balls remains at 295 yards.
  • H₀: μ = 295
  • H₁: μ ≠ 295
  • Sample results show a mean of 297.6 yards; if tests do not reject H₀, no adjustments are needed.

Testing for Population Proportions

  • General Test Statistic: z = (p - p₀) / √(p₀(1 - p₀) / n)
  • Rejection Rules based on critical values or p-values apply similarly as in means. Examples of studies often focus on increased proportions after interventions, like marketing promotions.

Relationship Between Hypothesis Testing and Confidence Intervals

  • Confidence intervals can provide insights into hypothesis tests.
  • If the hypothesized value lies within the interval, do not reject H₀; otherwise, reject it.

Conclusion

Understanding hypothesis testing involves recognizing how to structure hypotheses, the types of errors that can occur, and how to conduct analyses to draw valid conclusions. Practicing these concepts via examples and real-world cases is essential for mastery.