That could be, but that’s not how it works at some schools (I don’t know what Mich does).
Many schools use their predictive analytics model real time during the admission process. For example, the admission staff will craft a class, send the info to the enrollment management firm who runs number overnite…with the output being yield and/or revenue or whatever they want to measure. Then the admission staff adjusts the makeup of the class the next day, and rerun the model, and so on. Trinity college’s process is something like this and is discussed in a NYT article from last fall, as well as Paul Tough’s book The years that Matter Most.
Other schools put each application thru the predictive analytics model before the app is even read. The applications are deciled by likely yield, and applications from deciles with higher yields receive a more thorough read than those from applicants which the model says are less likely to attend if offered admission (models include variables that a school has found to be correlated with higher yield, etc)
I like Michigan, and am not being critical of them, I just thought that article was speaking out of both sides of their mouth.
I don’t know what the point would be of estimating yield after a school makes admissions decisions… perhaps they are doing different or enhanced marketing activities to those accepted kids they really want to attend, but leaving yield to chance after admission offers are made seems a bit backwards.