91Ó°ÊÓ

Hypothesis testing is a statistical technique used to decide whether there is enough evidence in a set of data to infer that a certain condition is true for the entire population. In our example of forest fires, we want to determine if the cause of fires depends on the region they occur in. This exercise begins with formulating two hypotheses:

• The null hypothesis (\(H_0\) ): Assumes there is no dependency between the cause of fire and the region, meaning they are independent.
• The alternative hypothesis (\(H_a\) ): Suggests that the cause of fire and the region are not independent, implying a relationship exists.

We test these hypotheses using a statistical method called the Chi-Square test. The goal is to determine whether any observed differences between data sets are due to chance or indicate a real variance.

The significance level, often denoted by alpha (\( \alpha \)), is a threshold set by the researcher before conducting hypothesis testing. It helps to decide whether to reject the null hypothesis. In this case, a 5% significance level (\( \alpha = 0.05 \)) is used.
A 5% significance level means that there is a 5% risk of concluding that a difference exists when there is none. This is called a Type I error. When the p-value obtained from our test is less than or equal to 0.05, the null hypothesis is rejected, indicating that the observed pattern is statistically significant.
Setting the significance level helps in making informed and reliable decisions based on the data, ensuring the results of the test are meaningful and not due to random fluctuations.

Degrees of freedom (df) are a crucial part of statistical tests, representing the number of values in the final calculation that are free to vary. They are determined based on the sample size and the nature of the data.
For the Chi-Square test, the degrees of freedom are calculated using the formula:
\[(df) = (number \ of \ rows - 1) \times (number \ of \ columns - 1)\]
In the forest fire example, there are 2 regions (rows) and 4 causes (columns). Therefore, the degrees of freedom are calculated as \((2-1) \times (4-1) = 3\).Knowing the degrees of freedom helps us determine the critical value from the Chi-Square distribution table. This critical value is used to compare against the calculated Chi-Square statistic to decide if we reject or fail to reject the null hypothesis.

Observed frequencies are the actual counts obtained from the data under investigation. In the context of the forest fire example, these frequencies show how often each cause occurred in both regions. On the other hand, expected frequencies represent what would be expected if the null hypothesis were true.
The expected frequency for any cell in the table is calculated by:
\[E_{i,j} = \frac{(\text{row total}_i \times \text{column total}_j)}{\text{grand total}}\]
For example, to find the expected frequency of arson-caused fires in Region A, the formula accounts for the proportion of total fires and applies it to specific categories.
The combination of comparing observed and expected frequencies is crucial in the Chi-Square test. It helps determine if deviations between observed and expected counts are significant or just due to random variations, ultimately guiding the hypothesis outcome.