Is There a Model Underlying the Mean?

The null hypothesis significance testing world, which gets maligned all the time, really does have a lot of similarity with he modelling. As we think about it, there is a model under there. Just like you said, the mean is a model. A te test has a model underlying it. So if you're talking about model based reasoning, i do think that a te test falls under that rubric. However, i also believe that testing a series of nested structural equation models using multi perimeter waldorf tests is also an effective way to look at data.