91Ó°ÊÓ

Statistical inference is a process that allows us to make conclusions about a population based on sample data. Imagine you have information on a small group and need to generalize it to a larger group. That's what inference helps with. In the context of hypothesis testing, it involves determining whether you can support a hypothesis about a population parameter, like a proportion or mean, based on sample data.

The steps in statistical inference often involve the following:

• Forming hypotheses: You develop a null and an alternative hypothesis to test.
• Gathering data: A sample is collected to be examined.
• Calculating statistics: Important metrics like sample proportion or mean are calculated.
• Making decisions: Based on calculated probabilities and test statistics, you decide whether to support or reject the hypothesis.

Each part ensures that you can make educated reasoning about the bigger picture beyond your sample.

The null hypothesis, often denoted as \( H_0 \), is a statement suggesting there is no effect or no difference, and it acts as a starting point for statistical testing. For our hotel example, \( H_0 \) posits that at least 90% of the hotels are full over the weekend. This presumption is made so that we can test it against the alternative hypothesis.

The null hypothesis is crucial because it establishes a default position that the evidence must contradict for the alternative hypothesis to be accepted. The goal of a hypothesis test is usually to challenge the null hypothesis by showing enough statistical evidence against it.

• Accepted until evidence suggests otherwise.
• Provides a baseline to compare against your data.
• Helps in measuring the degree of unusualness of your findings with respect to the null hypothesis.

Rejection of a null hypothesis suggests that the observed data significantly differ from what's expected under this initial assumption.

A P-value is a vital part of hypothesis testing and represents the probability of obtaining results as extreme as, or more extreme than, the observed results under the assumption that the null hypothesis is true. In essence, it indicates how likely your sample results are under the null hypothesis.

For instance, in our example, a P-value of approximately 0.0808 tells us the likelihood of seeing a sample proportion of hotels full as low as we observed—or even lower—if indeed 90% of hotels are supposed to be full. When the P-value is:

• Less than or equal to the significance level \( \alpha \): Strong evidence against the null, leading to its rejection.
• Greater than \( \alpha \): Not enough evidence to reject the null.

P-values offer a robust way to quantify the evidence against the null hypothesis, making them central to the decision-making process in hypothesis testing.

The test statistic is a standardized value that is calculated from sample data during a hypothesis test. It's used to determine the difference between your observed data and what was expected under the null hypothesis, using a mathematical formula.

For our scenario, the test statistic is calculated using the standard normal distribution since the sample is sufficiently large. Specifically, the statistic examines how far the sample proportion (0.8448) deviates from the hypothesized proportion (0.9) when adjusted for variations expected in random samples. This is derived through:

• \( z = \frac{\hat{p} - p_0}{\sqrt{\frac{p_0(1-p_0)}{n}}} \)
• Where \( \hat{p} \) is the sample proportion, \( p_0 \) the hypothesized proportion, and \( n \) the sample size.

The calculated test statistic, in this case, \( -1.40 \), helps compare the observed outcome directly against expected outcomes under the distribution theory, assisting in decisions about the null hypothesis.