91Ó°ÊÓ

In hypothesis testing, the null hypothesis is a fundamental concept. It's represented as \( H_0 \) and is the initial assumption or statement that there is no effect or no difference in a population parameter. For example, in a study investigating a population proportion \( p \), the null hypothesis may state that this proportion equals a specific value, like \( H_0: p = 0.45 \). The aim is to test whether there is enough statistical evidence in the sample data to reject this claim. Another way to think about the null hypothesis is as a statement of 'status quo' or no change. We use it as a baseline to compare against the sample data to see if any differences observed are due to random chance or if they're significant enough to result in a conclusion.

The alternative hypothesis is what you would consider to be true if the null hypothesis is proven false. Denoted by \( H_1 \), this hypothesis is essentially the opposite of the null hypothesis and is what the research aims to support. There are different forms of the alternative hypothesis:

• **Two-tailed**: This suggests that the parameter is not equal to the value specified in the null hypothesis (e.g., \( H_1: p eq 0.45 \)).
• **One-tailed**: Here, the parameter is presumed to be either less than or greater than the value in the null hypothesis (e.g., \( H_1: p < 0.72 \) or \( H_1: p > 0.30 \)).

The form of the alternative hypothesis affects the way we calculate critical values and, ultimately, how we make our decisions regarding the hypotheses.

The test statistic is a tool used in hypothesis testing to help in making decisions about the null hypothesis. Specifically, in the context of testing proportions, the test statistic is calculated as a Z-score. The formula for calculating this in the context of a proportion \( p \) is:
\[ Z = \frac{\hat{p} - p}{\sqrt{\frac{p(1 - p)}{n}}} \],
where \( \hat{p} \) is the observed sample proportion, \( p \) is the population proportion under the null hypothesis, and \( n \) is the sample size. This Z-score tells us how many standard deviations away our sample proportion is from the hypothesized population proportion. A large absolute value of the Z-score suggests a higher likelihood that the observed sample proportion is significantly different from the null hypothesis value, suggesting the null hypothesis might be rejected.

In hypothesis testing, the significance level, denoted as \( \alpha \), is essentially a threshold used to determine whether we reject the null hypothesis. Common values for \( \alpha \) are 0.05 or 0.01, and they represent the probability of making a Type I error—rejecting the null hypothesis when it's true. For instance, if \( \alpha = 0.05 \), there is a 5% risk of rejecting the null hypothesis mistakenly.
Depending on the alternative hypothesis, it can be a measure for two-tailed or one-tailed tests:

• In a **two-tailed test**, \( \alpha \) is split between the two tails of the distribution, focusing on the magnitude of deviation without direction.
• In a **one-tailed test**, all of \( \alpha \) is on one side of the distribution.

The significance level is used in setting the critical value, which in turn is a benchmark for determining whether the test statistic falls in a region where the null hypothesis is rejected.