How would you approach selecting the appropriate statistical test for a given dataset?

In my statistics class, I worked on a project analyzing survey data collected on student study habits. Our goal was to determine if there was a significant difference in study hours between students with high and low GPA. Since we had two groups and our data was normally distributed, I chose a t-test to compare the means. I outlined the steps: first, I verified that our data met the assumptions for a t-test, then I performed the test using statistical software. The result showed a significant difference, which we discussed in our presentation, emphasizing how choosing the correct test allowed us to draw valid conclusions about study habits among students.

During a health camp where I volunteered, we collected data on participant blood pressure levels before and after a nutrition workshop. I was tasked with analyzing this data to see if the workshop had any impact. Recognizing that we had paired data (same participants before and after), I chose to use a paired t-test. I explained the rationale to my team, emphasizing how this test accounts for the fact that the data points are related. After conducting the test, we found a statistically significant decrease in blood pressure, which we presented to our attendees, demonstrating the effectiveness of our workshop.

In my first job as a data analyst intern, I was involved in analyzing customer feedback ratings from our product surveys. We needed to determine if there was a difference in satisfaction ratings based on customer demographics. Understanding that the ratings were ordinal data, I opted for the Kruskal-Wallis test, as it allows for comparing more than two independent groups without assuming normal distribution. I communicated my approach to my supervisor, and after running the analysis, we discovered significant differences in satisfaction across demographics, which helped the marketing team tailor future campaigns.