05.11.2014 Views

national multiple family submetering and allocation billing program ...

national multiple family submetering and allocation billing program ...

national multiple family submetering and allocation billing program ...

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

method such as RUBS or <strong>submetering</strong>) compared to instances when residents did not receive a<br />

water bill, but the cost of water was included in the rent or homeowner’s dues (“in-rent”). Thus,<br />

“in-rent” was chosen as the comparison group.<br />

This type of variable transformation is necessary, because by definition there is not a<br />

numerical order to the categories of <strong>billing</strong> type; that is, if the number “1” is assigned to an inrent<br />

property, the number “2” to a RUBS property, <strong>and</strong> a “3” to a submetered property, it does<br />

not make sense to describe submetered properties as being 2 units of <strong>billing</strong> method greater than<br />

in-rent.<br />

To create the dummy variables for type of <strong>billing</strong> method, each property was classified<br />

using two variables, which were called RUBS <strong>and</strong> SUB. If residents at a property were billed<br />

using a RUBS methodology, the property was assigned a “1” for the RUBS variable <strong>and</strong> a “0”<br />

for the SUB variable. If residents at a property were billed using submeters, the property was<br />

assigned a “0” for the RUBS variable <strong>and</strong> a “1” for the SUB variable. Finally, if residents at a<br />

property did not receive a separate water bill from their rent or homeowner’s dues, the property<br />

was assigned a “0” for the RUBS variable <strong>and</strong> a “0” for the SUB variable. Thus, to fully classify<br />

<strong>billing</strong> method, both RUBS <strong>and</strong> SUB must be included in the model. The resulting B<br />

coefficients demonstrate the differences in average water use between RUBS properties vs. inrent<br />

properties <strong>and</strong> the differences in average water use between submetered properties vs. inrent<br />

properties. If one of these variables was not included in the model, the comparison would<br />

be between the variable included <strong>and</strong> all other property types; for example, if the SUB variable<br />

only was included, the B coefficient would represent the average water savings in submetered<br />

properties compared to in-rent <strong>and</strong> RUBS properties, adjusted for other factors included in the<br />

model.<br />

Many of the other variables found in one or more of the models presented below were<br />

included as dichotomous variables. For most, if the amenity or characteristic was present or<br />

“true” for a property, variable was coded as a “1” while the absence of the characteristic was<br />

coded as a “0.” These variables included: senior citizen/retirement community, play area, <strong>and</strong><br />

cooling tower. The B coefficient for these factors represents the difference in average amount of<br />

water observed in properties with the characteristics compared to those without, holding constant<br />

the other factors included in the model. “Property was built before 1995” was coded so that<br />

28 Prevalence of each characteristic can be found in Chapter 4 <strong>and</strong> Appendix B.<br />

153

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!