91Ó°ÊÓ

In hypothesis testing, we often want to determine if there is a specific directional effect. A **one-tailed test** is a statistical test where the area of interest is only on one side of the sampling distribution. In the Coca-Cola example, we hypothesized that the average consumption in Atlanta is greater than the national average. Consequently, our area of interest is only on the right side of the distribution.

When performing a one-tailed test:

• We state our null hypothesis (\( H_0 \)) as no effect or no difference, such as "the mean consumption is equal to the national average."
• Our alternative hypothesis (\( H_a \)) suggests a specific direction, e.g., "the mean consumption in Atlanta is greater than the national average."

To decide whether to reject the null hypothesis, we calculate a test statistic and compare it to a critical value specifically from one side of the t-distribution. This approach is more powerful when we are only interested in deviations in one direction, making conclusions like determining if consumption is strictly higher much more definitive.

The **significance level**, denoted by \( \alpha \), is a threshold set by the researcher before conducting a hypothesis test. It represents the probability of rejecting the null hypothesis when it is actually true, known as a Type I error. In our Coca-Cola example, the significance level was set to \( \alpha = 0.05 \).

What does \( \alpha = 0.05 \) mean?

• It indicates a 5% risk of concluding that the mean annual consumption in Atlanta is higher when it is not.
• This threshold helps determine the critical value from the t-distribution table.

Choosing the appropriate significance level is crucial. A smaller value means higher confidence in the results but increases the risk of missing a true effect (Type II error). It’s customary to use 0.05 in many research scenarios, serving as a balance between sensitivity and specificity.

The **t-distribution** is a probability distribution used in statistics, especially useful when dealing with sample data. Unlike the normal distribution, the t-distribution is employed when the sample size is small or when the population standard deviation is unknown, just as in our Coca-Cola example.

Key features of the t-distribution:

• It is symmetrical and bell-shaped, much like the normal distribution, but has heavier tails. This means there is more probability in the tails, providing a buffer for more extreme values in small samples.
• The shape of the t-distribution changes with degrees of freedom (df), which is \( n-1 \) for a sample size \( n \). The more data you have, the closer it resembles the normal distribution.

Using the t-distribution allows more accurate estimation of the critical values due to its adjustment for sample size-related variability. For our problem, with 35 degrees of freedom, the t-distribution informed us about the 1.690 threshold required for identifying if the test statistic was significant.