After reading about multicollinearity, I would definitively say that there was error in our regression models. Just look at the variables that most of us, if not all, used: Total Sq Ft, first floor area, bathrooms, bedrooms, total rooms, sale price, and sale year just to name a few. Each of these variables are closely correlated with at least one other variable, thus giving multicollinearity. Even when you first showed us how you started to analyze the data, Mr. Holman, you ran a correlation at the beginning and chose the variables that were closely related to the dependent variable, but some of the independent variables chosen were also closely related to one another. Not saying you were wrong what so ever, I mean obviously you knew/know what multicollinearity is. I am just using it as an example that from the get go, we had multicollinearity in our models.