Learn how multiple regression predicts college spending based on student enrollment and revenue. Drew College has $79 million in revenue and 2,200 students, but spends $91 million instead of the predicted $78 million, a $12 million gap that flags it as an outlier worth investigating.
Chapters
So let me address where the expected value comes from once, and we do it hundreds of times in this dashboard. So basically it comes from multiple regression analysis. All right, and on this axis we have total expenses, this axis we have students, number of FTE students, this axis total operating revenue. All right, so all this is, is a visual representation of what linear regression does. This is a multiple linear regression where the dependent variable is total expenses And then X1 is students, X2 is operating revenue. And what the regression will do is show you the average spending for any combination of students and operating revenue. And that average spending is that plane that it fits to the dataset there. Okay. So you may have noticed that the red dots here are Drew, and I can zoom in a little bit. Those red dots off the plane like that. That's Drew's problem that I'm about to talk about. They're off the plane. The plane is where they should be if they're going to be average for the peer set and given their revenue and students, and they're well above the plane. All right. And you can see it right here. Their operating revenue is $79 million. Um, their predicted spending given that revenue and their students, they're about 2,300, 2,200 students. Their predicted spending is $78 million. All right. Just below actually what their revenue is, which is exactly where you should be. All right. So that's this result up here. All right. So their expected spending is $79 million. Their actual is $91 million. The difference is $12 million.