91Ó°ÊÓ

Statistical hypothesis testing is a method used to make inferences about populations using sample data. It involves formulating two opposing hypotheses—the null hypothesis, denoted as \( H_0 \), and the alternative hypothesis, \( H_a \). The null hypothesis assumes that there is no effect or no difference, while the alternative hypothesis suggests that there is an effect or a difference.

When conducting a hypothesis test, the aim is to determine whether there is enough evidence in the sample data to reject the null hypothesis in favor of the alternative. To do this, a test statistic is calculated from the sample data. This statistic helps us decide whether the observed data is consistent with the null hypothesis or if it supports the alternative hypothesis.

• The test statistic is compared to a critical value derived from a theoretical distribution, which could be normal, t, chi-square, etc., depending on the test type.
• If the test statistic is extreme in the direction specified by the alternative hypothesis, \( H_0 \) is rejected.
• This process involves determining a significance level (commonly 0.05), below which the null hypothesis is rejected.

In goodness-of-fit tests, the goal is to assess how well the sample data adheres to a specified distribution by evaluating the discrepancy between observed and expected frequencies.

The chi-square distribution is a key component in goodness-of-fit tests and other statistical tests like the test for independence. It is defined only for non-negative values and is skewed to the right, especially when the degrees of freedom are low. This type of distribution is very useful in evaluating how observed frequencies deviate from expected frequencies.

The chi-square distribution is characterized by its degrees of freedom, which affect its shape. The degrees of freedom in a goodness-of-fit test are usually calculated as the number of categories minus one. More degrees of freedom make the distribution more symmetrical.

• In goodness-of-fit tests, the test statistic follows a chi-square distribution under the null hypothesis.
• As the sample size increases, the chi-square distribution approaches a normal distribution.

Because we assess the likelihood of observing a value as extreme or more extreme than the test statistic, the right tail of the chi-square distribution is of particular interest. Large values in this tail indicate that our observed data significantly deviate from what was expected, thereby suggesting that the null hypothesis might not be valid.

In the context of goodness-of-fit tests, interpreting the test statistic involves determining how well the observed data fits the theoretical distribution. This is primarily determined using the chi-square test statistic, computed as \[X^2 = \sum \frac{(O_i - E_i)^2}{E_i}\]where \(O_i\) is the observed frequency, and \(E_i\) is the expected frequency for each category.

The resulting test statistic \(X^2\) is compared against a critical value from the chi-square distribution table. The interpretation is straightforward:

• If \(X^2\) is larger than the critical value at a given significance level, it suggests a significant discrepancy between observed and expected data, prompting us to reject \( H_0 \).
• If \(X^2\) is less than the critical value, there is insufficient evidence to reject \( H_0 \), indicating that the observed data fits the expected distribution well.

The emphasis on the right tail for this test means we're looking for large values of \( X^2 \) to determine if the observed frequencies differ significantly from what's expected under the null hypothesis. This inherent focus on larger discrepancies helps explain why goodness-of-fit tests are right-tailed.