Content area
Full Text
When deriving a value estimate, we appraisers may find it insufficient to rely solely on paired sales analysis or on developing a sample of arrayed comparable sales and calculating the standard deviation. Although good, it may not seem great. Greater statistical precision or inference may be offered using one characteristic of the comparable sales and correlating it to respective sale prices of the comparables. An alternative to this is linear regression analysis, which provides more substantial evidence of market-derived adjustments than using only two sales.
Is expensive, complicated computer software needed? Absolutely not, just use your trusty HP-12C calculator. We will use this inexpensive hand-held calculator to demonstrate the use of simple linear regression in deriving market-based adjustments.
This example includes all the keystrokes on the calculator in order to assist those who have not previously used regression techniques. Technical jargon has been avoided as much as possible.
The example uses simple linear regression to indicate the amount of adjustment for a selected physical characteristic or element. We need to assume that the only difference between the comparable sales and the subject property is one characteristic and that all other differences have been accounted for. In other words, no other adjustments apply. Characteristics may include but are not limited to square feet, acres, gross income, number of sites, or number of apartments. All characteristics are relative to sale price. Simple linear regression analysis attempts to reflect such a relationship, if one exists, and to what extent. We need to bear in mind also that statistical inference may be limited to the degree that it may not incorporate sufficient appraiser judgment.
In the example used, we will imagine...