In this presentation, we explore the crucial role of predicion, confidence, and tolerance limits. Given that they are easy to calculate and appear in most textbooks, they are commonly used; however, that also means that they are often misused or misinterpreted. One of the current issues lies in accurately estimating these limits for mixed models, which is a serious limitation in most industries. 

While JMP uses the best option available to estimate these limits (based on a paper published in 2019 in the journal, Wiley Medicine), we can run into trouble if we're not careful, as the method does have some shortcomings. As a consulting company, the methods we use need to be solid. Therefore, we decided to delve deeper into the issue and evaluate these limits through simulation in JMP. 

The presentation focuses on the storytelling of the problem, making people across industries aware of these statistical dilemmas, along with a demonstration of how JMP was used to simulate different scenarios and how models were evaluated. It  showcases a graphical interface made for this purpose, as well as model building and producing an interpretable output.