91Ó°ÊÓ

In hypothesis testing, the significance level is a critical component that defines how extreme the data must be to reject the null hypothesis. You can think of the significance level, denoted as \( \alpha \), as a threshold percentage. Typically, common values are 0.05, 0.01, or even 0.10, but it's set at 0.01 in our exercise. This means we're willing to accept a 1% chance of incorrectly rejecting the null hypothesis when it's actually true (this is known as a Type I error).
A lower significance level indicates a more stringent criterion for making a decision. In our exercise, using a 1% significance level means that we need strong evidence against the null hypothesis to decide that the pollution index differs between Denver and Englewood. Simply put, the more remarkable the findings must be to challenge the null hypothesis at such a low \( \alpha \).

• \(\alpha = 0.01\) implies extremely strong evidence is needed to claim a difference.
• Significance levels help frame how confident we need to be in order to reject \( H_0 \).

The test statistic is a standardized value that is calculated from sample data during a hypothesis test. It is used to decide whether to support or reject the null hypothesis. Our specific situation involves comparing means with known standard deviations. Given this, we use a z-score as our test statistic.
In our exercise, the formula to compute the z-score is:\[z = \frac{\bar{x}_1 - \bar{x}_2}{\sqrt{\frac{\sigma_1^2}{n_1} + \frac{\sigma_2^2}{n_2}}} \]The numerator \( (\bar{x}_1 - \bar{x}_2) \) represents the observed difference in sample means, while the denominator adjusts this difference for the sample variability.
Once calculated, this z-score tells us how many standard deviations our observed difference is from the null hypothesis's assumption (here, that the means are equal). In our example, a z-score of approximately 0.949 suggests the observed difference isn't exceptionally unusual, given the null hypothesis. This forms a critical part of deciding whether to reject or retain \( H_0 \).

• Used to compare observed data to what's expected under \( H_0 \).
• A higher absolute z-score means the data is more unusual under \( H_0 \).

The null hypothesis \((H_0)\) is a starting assumption for statistical testing. It proposes that there is no effect or difference in a particular context. In terms of our exercise, \(H_0\) states that the mean pollution indices for Denver and Englewood are equal, or mathematically, \(\mu_1 = \mu_2\).
This hypothesis serves as a default statement that the data has no compelling effect or pattern to observe. It remains our reference point throughout the testing process, unless our calculations provide sufficient evidence to conclude otherwise. It's crucial here because all of the test statistic's interpretations hinge on our understanding of this \( H_0 \).

• \(H_0: \mu_1 = \mu_2\) suggests no difference in real terms.
• This is the hypothesis we aim to test and potentially reject.

The alternative hypothesis \((H_a)\) suggests that a pattern or difference does exist and contradicts the null hypothesis. In our example, \(H_a:\mu_1 eq \mu_2\) posits that the pollution levels in Denver and Englewood differ, without specifying which is higher or lower.
This alternative is crucial as it is the hypothesis we look to support with our data. It provides an insight or conclusion we're trying to prove. If statistical analysis (test statistic) supports the evidence that contradicts \( H_0 \), we turn to \( H_a \) as the likely reality.
In practical applications, considering \( H_a \) helps provide the framework for creating test statistics and calculating probabilities that ultimately confirm (or deny) significant differences.

• \(H_a: \mu_1 eq \mu_2\) allows for any substantial difference.
• It's what we wish to prove in opposition to \( H_0 \).