When to use a directional vs. non-directional hypothesis test in research

📚 Quick Study Guide

• 🎯 Directional (One-tailed) Hypothesis Test: Use when your research predicts the direction of the effect (increase or decrease). For example, "Drug X will *increase* reaction time."
  • 📊 Hypothesis: $H_0: \mu \le 0$, $H_1: \mu > 0$ (right-tailed) OR $H_0: \mu \ge 0$, $H_1: \mu < 0$ (left-tailed)
  • 🧪 Critical Region: Located on one side of the distribution.
• ⚖️ Non-Directional (Two-tailed) Hypothesis Test: Use when your research only predicts an effect, but not its direction. For example, "Drug X will *affect* reaction time."
  • 🔢 Hypothesis: $H_0: \mu = 0$, $H_1: \mu \neq 0$
  • 🌍 Critical Region: Split between both tails of the distribution.
• 💡 Key Difference: Directional tests have more power to detect an effect if the direction is correct, but no power if the effect is in the opposite direction. Non-directional tests are more conservative but can detect effects in either direction.
• 🤔 Decision Rule:
  • Directional: Reject $H_0$ if test statistic falls in the single critical region.
  • Non-Directional: Reject $H_0$ if test statistic falls in either of the two critical regions.
• ✍️ P-value:
  • Directional: Use the p-value directly from the statistical test if it matches the hypothesized direction.
  • Non-Directional: Double the p-value from the statistical test.

Practice Quiz

1. Which of the following scenarios calls for a directional hypothesis test?
   1. (A) Investigating whether a new fertilizer affects plant growth.
   2. (B) Determining if a drug alters cognitive function.
   3. (C) Testing if a training program *increases* employee productivity.
   4. (D) Examining the relationship between age and income.
2. When should you use a non-directional hypothesis test?
   1. (A) When you predict a specific increase in a variable.
   2. (B) When you expect a specific decrease in a variable.
   3. (C) When you are unsure of the direction of the effect.
   4. (D) When you want to increase the power of your test.
3. In a directional (right-tailed) test, the critical region is located:
   1. (A) On both tails of the distribution.
   2. (B) On the left tail of the distribution.
   3. (C) On the right tail of the distribution.
   4. (D) In the center of the distribution.
4. What is the primary difference between one-tailed and two-tailed tests regarding the p-value?
   1. (A) One-tailed tests always have smaller p-values.
   2. (B) Two-tailed tests always have smaller p-values.
   3. (C) The p-value is doubled in two-tailed tests.
   4. (D) The p-value is halved in one-tailed tests.
5. If you hypothesize that a new teaching method will *decrease* test scores, which type of test should you use?
   1. (A) Two-tailed test.
   2. (B) One-tailed test.
   3. (C) Either one-tailed or two-tailed, depending on sample size.
   4. (D) A non-parametric test.
6. Which of the following is a scenario where a two-tailed test is most appropriate?
   1. (A) Testing if a drug reduces blood pressure.
   2. (B) Testing if a new diet increases weight.
   3. (C) Testing if a marketing campaign has *any* impact on sales.
   4. (D) Testing if a new fertilizer increases crop yield.
7. Why is it important to decide between a directional and non-directional test *before* analyzing the data?
   1. (A) To increase the sample size.
   2. (B) To avoid bias and ensure valid results.
   3. (C) To simplify the calculations.
   4. (D) It doesn't matter when you decide.