Describe a situation where using a non-parametric test would be more appropriate than a parametric test?

In my final year at university, I conducted a research project analyzing survey results from a group of students about their study habits. The data collected was ordinal, as students rated their study habits on a scale from 'very poor' to 'excellent.' Because the data was not normally distributed and did not meet the assumptions of parametric tests, I opted for the Mann-Whitney U test to compare the study habits of two groups: those who studied more than 10 hours a week versus those who studied less. This choice allowed me to draw meaningful conclusions about the differences without being misled by the data's distribution.

During my time volunteering for a local nonprofit, I helped collect feedback from community members about a new initiative. The feedback was gathered through a ranking system, which resulted in ordinal data. To analyze the differences in satisfaction levels between various age groups, I chose the Kruskal-Wallis test instead of a one-way ANOVA because the sample sizes were small and the data was not normally distributed. This decision helped us identify which age group was most satisfied, guiding future project development effectively.

In my first job as a research assistant, I was involved in evaluating employee performance based on peer reviews that were collected on a Likert scale. Recognizing that the data were ordinal and not normally distributed, I decided to use the Wilcoxon signed-rank test to analyze differences in performance before and after a training program. This method provided valuable insights into the effectiveness of the training, which helped improve our future initiatives.