91Ó°ÊÓ

Hypothesis testing is a critical component of statistical analysis. It helps us determine if there is enough evidence to support a particular claim about a dataset. In our exercise, we need to test if the success rate of percutaneous nephrolithotomy (PN) is independent of the size of kidney stones.
To begin, we set up two hypotheses:

• The Null Hypothesis (\(H_0\)) - This assumes that the success of PN is independent of stone size.
• The Alternative Hypothesis (\(H_a\)) - This assumes that the success of PN depends on the size of the stone.

We will use a chi-square test for independence to examine these hypotheses. If the data provides sufficient evidence against the null hypothesis, we will reject it in favor of the alternative hypothesis. It's important to set a significance level, \(\alpha\), which is the probability of rejecting the null hypothesis when it is true. In this exercise, we use \(\alpha = 0.05\). This means there is a 5% risk of concluding that the stone size affects the success rate when it actually does not.

A contingency table is a data matrix that displays the frequency distribution of the variables, in our case, the success of PN and stone size. It allows us to visually compare the observed data with what might be expected if, indeed, the two variables were independent.
In the exercise, we created the following table based on the data: - For stones of size <2 cm, there are 234 successes and 36 failures out of 270 trials.- For stones of size ≥2 cm, the table shows 55 successes and 25 failures out of 80 trials.
The contingency table helps us structure our calculations. Specifically, it will be key for finding expected frequencies for each cell, which we'll need to carry out the chi-square test. Expected frequencies are based on the assumption that the null hypothesis is true. They are calculated as: \[E_{ij} = \frac{(\text{Row total})_i \times (\text{Column total})_j}{\text{Grand total}}\]where \(E_{ij}\) represents the expected frequency for cell \(i, j\). This is crucial, as large deviations between observed and expected frequencies signal potential dependence between our variables.

The p-value is a probability measure that underpins hypothesis testing. It informs us about the likelihood of observing our data if the null hypothesis is true.
In the chi-square test for independence, after calculating your \(\chi^2\) statistic, you use a chi-square distribution table to find your p-value. Here’s how it works:

• First, you calculate the \(\chi^2\) statistic using: \[\chi^2 = \sum \frac{(O_i - E_i)^2}{E_i}\]where \(O_i\) is the observed frequency and \(E_i\) is the expected frequency.
• Once you have your \(\chi^2\) value, compare it to a chi-square distribution table with your degrees of freedom (calculated as \((\text{rows}-1)(\text{columns}-1)\)).
• Determine your p-value from the table or using a calculator. This step tells you the probability of obtaining a test result at least as extreme as the one observed, given that the null hypothesis is true.

If the p-value is less than the \(0.05\) significance level, it suggests a significant departure from the null hypothesis, thus leading you to reject it. Conversely, a higher p-value implies insufficient evidence against the null hypothesis.