Yes you can. Sometimes attributes are not marked as significant because we do not see enough variation in the data at the time of modeling. However, we encourage appraisers to review attributes again in order to determine if they should make adjustments. This is where an appraiser’s expert market knowledge is very valuable. One word of caution, though, on double-counting. Sometimes two attributes (like bedrooms and bathrooms) are highly correlated. In that situation, it is often the case that the regression chooses one of those attributes to give most of the adjustment value. In a situation of high collinearity, it is important for the appraiser to avoid making adjustments for bedrooms (which may be insignificant) when the value adjustments for bedrooms are mostly captured by bathrooms.
Articles in this section
- FAQ Regression - What does R squared signify?
- FAQ Regression - How can bedrooms be considered an insignificant attribute?
- FAQ Regression - What is collinearity? Why is it important?
- FAQ Regression - Do appraisers have to use the regression results?
- FAQ Regression - Can I make an adjustment even when an attribute is not considered significant?
- FAQ Regression - What’s the value of analyzing 2000 properties? Don't we just need three?
- FAQ Regression - What does significant mean?
- FAQ Regression - When should I make adjustments to attributes other than the ones HouseCanary analyzed?
- FAQ Regression - What do the sliders mean?