91Ó°ÊÓ

When we perform hypothesis testing, we begin by stating two contradictory claims about a population parameter. The null hypothesis (\( H_0 \)) represents the status quo or the baseline that we assume to be true before we collect any data. For instance, in our example, the null hypothesis is that the new distribution method (with an eye patch) doesn’t affect the questionnaire return rate, which remains at 40%. In formal terms, that's written as \( H_0: p = 0.4 \
\
\).
On the flip side, the alternative hypothesis (\( H_1 \) or \( H_a \)) represents a new theory or claim that we aim to support, typically indicating that there is an effect or difference. In the given problem, the researcher posits that the response rate will increase thanks to the stigmatization associated with the eye patch. Mathematically, this is stated as \( H_1: p > 0.4 \
\
\). It's important to choose the hypotheses carefully and ensure they're mutually exclusive; in other words, if one is true, the other must be false. A good hypothesis test aims to gather enough evidence to reject the null hypothesis in favor of the alternative hypothesis.

The sample proportion, often denoted as \( p\text{-hat} \), is a very useful statistic in hypothesis testing, especially when the underlying question revolves around proportions or likelihoods within a population. It is calculated by dividing the number of 'successes' observed in the sample by the total number of observations in the sample.

For the example under discussion, the success is receiving a returned questionnaire; thus, out of the 200 distributed, 109 were returned, resulting in a sample proportion: \( p\text{-hat} = \frac{109}{200} = 0.545 \). This empirical evidence from the sample is then utilized to analyze the likelihood of the null hypothesis being true. If the sample proportion greatly differs from the null hypothesis value, it can lead to the rejection of the null hypothesis.

In the realm of hypothesis testing, the test statistic is a critical value that allows us to decide whether our sample data provides enough evidence to reject the null hypothesis. It’s essentially a standardized value that measures the degree of deviation of the sample statistic from the null hypothesis.

In the context of our example that deals with proportions, the test statistic can be a Z-score, which is calculated using this formula: \( Z = \frac{p\text{-hat} - p_0}{\sqrt{\frac{p_0 * (1 - p_0)}{n}}} \), where \( p_0 \) represents the proportion as per the null hypothesis, \( n \) is the sample size, and \( p\text{-hat} \) is the sample proportion. After calculations, a Z-score of 4.52 indicates a significant difference from the hypothesized population proportion. The magnitude of the test statistic plays a pivotal role in determining the probability of observing our sample result (or something more extreme) if the null hypothesis were true.

The significance level, denoted by \( \alpha \), is a threshold that we set for deciding whether to reject the null hypothesis. It’s the probability of making a Type I error, which occurs if we incorrectly reject a true null hypothesis. Common standard significance levels include 0.05, 0.01, and 0.10.

In hypothesis testing, once we've calculated the test statistic, we compare it against critical values corresponding to the chosen significance level. These critical values are the cutoff points that determine the boundaries of the regions where we would reject \( H_0 \). If our test statistic falls into this rejection region, we have sufficient evidence to reject the null hypothesis. In our eye patch questionnaire example, by using \( \alpha = 0.05 \) for a one-sided test, the critical Z-value is approximately 1.645. Since our calculated Z-value of 4.52 exceeds this, we reject the null hypothesis, implying that the data does suggest – with a high level of confidence – an increase in the response rate.