91Ó°ÊÓ

In hypothesis testing, setting up null and alternative hypotheses is the foundational step to take before conducting any test. These hypotheses help us decide whether there is enough evidence to support a particular claim about a population parameter.

For the null hypothesis (denoted as \( H_0 \)), we assume that the statement being tested is true until evidence suggests otherwise. It's often a statement of "no effect" or "no difference." In our exercise, the null hypothesis is that the proportion of female wolves (\( p \)), is 0.5, meaning there's no change from the historical proportion.

Conversely, the alternative hypothesis (denoted as \( H_a \)) challenges the null hypothesis. It's what you want to prove. Here, the alternative hypothesis suggests that the proportion of female wolves is now less than 0.5, indicating a decline in the population proportion.

These hypotheses can be summarized as:

• \( H_0: p = 0.5 \) (Null hypothesis)
• \( H_a: p < 0.5 \) (Alternative hypothesis)

This logical framework allows us to systematically test and compare statistical evidence.

A Z-test for proportions is a statistical method used to determine if there's a significant difference between the observed sample proportion and a theoretical population proportion. In this scenario, it helps determine if the percentage of female wolves has changed over time.

The process involves calculating the sample proportion, \( \hat{p} \), which is simply the ratio of the number of observed females to the total sample size. In the given exercise, we have 10 females out of 34 wolves, giving us a sample proportion of \( \hat{p} \approx 0.2941 \).

To perform the test, calculate the Z-test statistic using the formula:

\[ z = \frac{\hat{p} - p_0}{\sqrt{\frac{p_0(1-p_0)}{n}}} \]

where \( p_0 \) is the hypothesized population proportion (0.5 in this case), \( n \) is the sample size, and \( \hat{p} \) is the sample proportion. This calculation provides a Z-score, which allows us to determine the position of the sample proportion within the context of the normal distribution curve.

• If the Z-score is extreme (far from zero), this suggests that the sample proportion is significantly different from the hypothesized population proportion.

This method is useful when sample sizes are large enough (generally \( n > 30 \)) to justify the approximation to the normal distribution.

The significance level, denoted as \( \alpha \), is a threshold set by the researcher to determine how much evidence is necessary to reject the null hypothesis. It's a way to gauge the risk of incorrectly rejecting a true null hypothesis, known as a Type I error.

In hypothesis testing, the significance level defines the probability of making such an error. Common choices for \( \alpha \) are 0.05, 0.01, and 0.10, but it can be any value depending on the researcher's tolerance for risk.

In the example we examine, the significance level is set at 0.01. This is a strict level, meaning the evidence must be strong enough to reject \( H_0 \), as only 1% chance of an error is tolerated. A lower \( \alpha \) reflects a need for more convincing data to support the alternative hypothesis.

When interpreting results, compare the calculated Z-score to the critical value at this significance level. If the Z-score falls in the critical region (defined by \( \alpha \)), it indicates sufficient evidence to reject the null hypothesis. Thus:

• If the test statistic is less extreme than the critical value, \( H_0 \) is not rejected.
• If the test statistic is more extreme, \( H_0 \) is rejected in favor of \( H_a \).

This structured approach helps clarify whether observed differences are genuine or due to random chance.