Alliance for Lighting InformationAlliance for Lighting InformationThe notion that the amount of outdoor lighting can be associated with the amount of commercial activity on a geographic area is unfounded especially when the geographic areas are defined in an inconsistent manner such as are zip codes. Zip codes do not all cover the same amount of physical area, do not all have the same population density, nor do they tell us anything about the physical make up of the area - urban versus suburban vs. industrial vs. service oriented. Areas of high commerce can have drastically different amounts of exterior lighting. Zip codes encompassing areas of downtown Los Angeles and San Francisco could have relatively high commercial activity, but these areas have low amounts of exterior lighting as parking is usually provided in parking structures under the buildings, whereas more outlying zip codes may have equally high commerce but provide much more exterior lighting in the form of surface parking or landscaping lighting as land is not as precious a commodity.
The sample selection is further muddied by the notion that business activity (commerce?) can be related to the number of employees in an area and that there is some relation to the number of employees and the amount of exterior lighting. Better, more accurate and more appropriate data might have been collected if the sample selection process had been based on zoning districts rather than on zip codes and number of employees. Since the final recommendations of this report attempt to assess the impact of the proposed lighting power densities (watts per square foot) on the state's energy consumption, would it not have been more sensible to include the actual area (in square feet) of the state covered by the different 'functional use areas' in the scale up?
As examples of how the utilized system falls apart, the utilized system implied that an appropriate sample size for gas stations was 7 although only 5 can really be considered since 2 of the stations surveyed are missing part of their data set. The system also implied that an adequate sample size for grocery stores is 3 (only 2 sites have illuminance measurements for the parking lots), the adequate sample size for hospitals is 8 (7 with illuminance data), the adequate sample size for exterior outdoor sales is 4 (only 2 have illuminance data), and the adequate sample size for full service restaurants is 7. From the same 1997 Zip Code Business Patterns data base used as the basis for this report the following numbers were obtained: there are 8721 Gas Stations, 7281 Supermarkets, 669 Hospitals and HMO Centers, 3340 Auto Dealers, and 22871 Full Service Restaurants in the state of California. The sample sizes implied by the chosen methodology do not seem to be adequately modeling the exterior lighting population. Since Auto Dealers, Gas Stations and Supermarkets are arguably some of the largest users of exterior lighting, the errors introduced to the 'scaled up data' due to such small relative sample sizes could be quite significant.
In addition, the list of zip codes used indicates that the surveyors visited few, if any, sites in area where they themselves did not feel comfortable. This introduces a cultural and socio-economic bias into the data.
In addition, this analysis does not include any public right of way lighting. There is no consideration given to either the amount of energy consumed or the impact of roadway lighting. This report indirectly implies that the whole energy problem is strictly the result of private sector exterior lighting and that no benefit would be had from addressing poorly lighted roadways.
A comprehensive publication on statistical methods can be found at: http://www.sportsci.org/resource/stats/
This site is provided as a public service by resodance publishing co. and no guarantee is implied or provided.
This page is served & maintained by &
last changed on 7 May 03 by
