11/01/88 - Laserfiche WebLink

Home

The method used to determine the actual relationship between variables and ~rice is called regression analysis. Multiple regression analysis is used' to sort a large number of possible variables into an order according to statistical significance and importance. In this research, multiple regression analysis was performed to sort the relationships between 15 housing variables and sale price for 1812 homes within five miles of three landfill sites {Anoka Regional Sanitary Landfill in Ramsey, Waste Disposal Engineering in Andover and Oak Grove Landfill in Oak Grove} in Anoka County. Each landfill was considered separately. One can readily imagine a situation in which several housing variables could be found to influence significantly the sale price of a house. If one wished to know the marginal difference in price that could be attributed to each significant variable, one must calculate the Beta weight for each variable. The Beta weight is expressed as (B=$x) and shows the dollar value added or subtracted from price by a unit change (+/- 1) in a significant variable. For example, let us assume that the number of square feet in a house had a significant affect on the final price. Beta weights will tell us the change in price that is achieved if one changes the number of sqc~are feet. In this example, let us2further assume the Beta weight for, ft is 20. If one house has 100 ft more than an otherwise simila~ house, ~, should be worth $2000 more. The calculation is simply $20/ft' x 100 ft~ = $2,000. The purpose of analytical methods like the Model of Hedonic Value is to isolate characteristics (variables) which can serve as reliable indicators of some phenomenon (in this case, price of a house). It is assumed that significant variables account for differences in price. It is further assumed that marginal differences in price can be attributed to degrees of change in significant variables. Many variables used in previous Hedonic model studies have been used to test differences in housing prices in a variety of geographic locations. Sometimes, however, researchers add variables to the conventional model which they feel may be important when they wish to measure the impact of certain forces on value -- like the building of a landfill, for example. Once researcherd habr assembled a list of possible variables to measure and has considered the significance of these variables in determining variations in housing price, they will want to know if the variables he has cl~osen are the right ones. The statistical test for this is expressed as (R-). The R2 value shows the percent of variance in price which can be explained by all the independent variables in the model, taken together. In a reliable Hedonic Model, one would want to know that the variables one has selected to study, lumped together, could explain as much of