91Ó°ÊÓ

Hypotheses testing involves making an assumption about a parameter and finding evidence to support or refute this assumption with data. We start by formulating a **null hypothesis** (\(H_0\)) and an **alternative hypothesis** (\(H_a\)). In the context of an ANOVA test, the null hypothesis states that all group means are equal, implying no effect or no difference. For the toy store price example, this means:
\(H_0: \mu_1 = \mu_2 = \mu_3\), where \(\mu_1, \mu_2,\) and \(\mu_3\) represent the average prices of toys in discount, variety, and department stores respectively.
The alternative hypothesis suggests that at least one group mean is different, indicating a potential difference:
\(H_a:\) at least one \(\mu\) is different.
This method helps in decision-making, allowing us to evaluate if observed data is consistent with the null hypothesis or if there is enough evidence to support the alternative.

The **significance level** (\(\alpha\)) in hypothesis testing determines the threshold for rejecting the null hypothesis. It is the probability of falsely rejecting a true null hypothesis (Type I error). In most statistical analysis, a common significance level is 0.05, as used in this toy price example.
Using \(\alpha = 0.05\) means there is a 5% risk of concluding that a difference exists when in fact, it does not. It's like setting a bar for the evidence needed to reject \(H_0\). If the probability that the data supporting \(H_a\) happening by random chance is lower than 5%, we reject \(H_0\).
This threshold ensures that critical decisions are based on a strong likelihood of correctness rather than arbitrary results.

**Statistical analysis** of data involves organizing, interpreting, and drawing conclusions. The ANOVA test is a powerful statistical technique used to compare means among three or more groups. ANOVA helps assess whether the observed differences between group means are more than what we might expect by chance.
For this test, we calculate:

• Group means: the average price of toys for each store type.
• Overall mean: the average price of toys across all stores.
• Sum of squares between groups (SSB): measures variability between different groups.
• Sum of squares within groups (SSW): measures variability within each group.

The resulting calculations break down into variance components:

• Mean Square Between (MSB) = SSB divided by its degrees of freedom.
• Mean Square Within (MSW) = SSW divided by its degrees of freedom.

This analysis prepares the data for the calculation of the F-ratio, determining if group means are statistically different.

The **F-ratio** is a critical part of the ANOVA test and represents the ratio of variation between group means to the variation within the groups. This ratio helps in determining if the group means differences are statistically significant.
The computation involves:

• F-ratio = MSB/MSW

The higher the F-ratio, the greater the likelihood that the observed data groups have different means.
We assess the F-ratio against a critical value from the F-distribution table, which considers the degrees of freedom for between groups (dfb) and within groups (dfw). In our example: dfb=2 and dfw=12.
If the F-ratio exceeds the critical value from the F-table (at \(\alpha = 0.05\)), the null hypothesis is rejected, concluding a statistically significant difference in toy prices among different store types. This approach remains foundational for comparing multiple group means efficiently.