91Ó°ÊÓ

A non-parametric test is a type of statistic used when the data doesn't fit well into the normal distribution assumption. These tests are useful when data ranks are being compared rather than actual values. Unlike parametric tests which require interval or ratio scale data that follow a normal distribution, non-parametric tests make fewer assumptions about the data and can handle ordinal data or non-normal interval data effectively.
In the context of this exercise, we are using the Wilcoxon rank-sum test, a common non-parametric test that does not assume normal distribution of the data. This test is used to determine if there are differences in the ranks of two independent groups, in this case, the flight delay times from Delta and Alaska airlines. Hence, it is perfectly suited for rank-based comparisons and situations where sample sizes are small or the data contains outliers.

• No assumptions about data distribution
• Useful for ordinal or skewed data
• Compares median values across groups
• Appropriate for small sample sizes

On-time performance in aviation involves evaluating how efficiently an airline manages to adhere to its schedule, particularly its arrival and departure times. Travelers prefer on-time flights as it minimizes waiting time and delays, ensuring smoother travel experiences.
In this exercise, on-time performance is being assessed by looking at the delay times for each flight on Delta and Alaska airlines. Shorter delays suggest better on-time performance. By comparing the delay times of these airlines, the traveler aims to switch to the airline that consistently arrives on time or with the least amount of delay.
This assessment is essential for decision-making, especially for frequent travelers like in the given scenario, who may opt for the airline with the superior on-time performance.

• Evaluates punctuality of flight schedules
• Critical for customer satisfaction
• Aids decision-making for frequent travelers

Statistical significance is used to determine if the observed effect or outcome seen in the data is likely due to something other than random chance. It helps confirm whether the findings from an experiment or study are reliable and not just happenstance.
In the Wilcoxon rank-sum test, statistical significance is evaluated by comparing the calculated test statistic to a critical value from statistical tables, which corresponds to the chosen significance level (alpha, \( \alpha \)). In this exercise, a significance level of \( \alpha = 0.01 \) indicates a 1% risk of concluding that there is a difference in on-time performance between the airlines when there is actually none.
If the test statistic is less than or equal to the critical value, the result is deemed statistically significant, suggesting a genuine difference in performance between Delta and Alaska, beyond random variation.

• Determines reliability of study findings
• Uses a significance level (alpha) to assess risk
• 0.01 significance level in this context implies strong evidence needed to conclude a difference

Rank assignment is a key step in non-parametric testing where each observation in the combined dataset is given a rank based on its value, usually from smallest to largest. In the presence of ties, average ranks are assigned to each tied value.
In this exercise, the process involves ranking the combined flight delay times for Delta and Alaska airlines. Assigning ranks helps transform the data into an ordinal scale, which is suitable for non-parametric analysis. This method allows the reduction of the impact of outliers and focuses more on median values.
The ranks are then used to compute the rank sums for each group (Delta and Alaska), which are pivotal for conducting the Wilcoxon rank-sum test. These ranks determine the test statistic which is compared against critical values for assessing statistical significance.

• Converts measurements to ordinal data
• Focuses on data ranks rather than exact values
• Handles ties by assigning average ranks
• Enables calculation of rank sums for non-parametric testing