Good question and the answer is, “it depends.” There are instances where a team is testing their process to see if it’s performing to standards. If the standard is something like, “Average of 5 minutes response time”, then the team would take one data set that represents the current state of response times and compare it to the standard. They are checking to see if there’s a difference between the current state and the standard. If you’re wondering about the required sample sizes in each data set, that’s another topic. There are sample size calculators in most statistical packages as well that help determine valid sample sizes. That would happen prior to data collection. I hope that’s helpful!