There are three kinds of lies: lies, damned lies, and statistics.
Mark Twain, paraphrasing Benjamin Disraeli
We have a new and very interesting recent report on green building to examine, the Regional Green Building Case Study Project: A post-occupancy study of LEED projects in Illinois . The Illinois report studied a mix of projects of various certifications levels, certified under various versions of LEED, with various applications, that used various baselines, and that used various reporting methods for utilities. The small sample and disparate projects involved lends itself towards a scattershot of data.
The study found a range of cost premiums for green construction (after grants and incentives) at 0.6% to 6.9% of project cost with 3.8% increase as the mean. The Illinois report found that projects focusing on energy efficiency tended to have better energy efficiency (quite a V-8 moment I am sure). Interestingly, the Illinois report found no meaningful correlation between the level of LEED certification and the energy performance of the buildings involved … this may be an off-shoot of the limited sample size.
The most interesting point from the Illinois study was in the details regarding energy modeling. Sixteen projects used energy design models. Of these sixteen projects, twelve projects performed worse than design models … six of them using 100% to a whopping 250% more energy than anticipated in the model . Seventeen projects used at least baseline models and remarkably seven of the projects showed actual energy use above the baseline!
To me, to have energy use of these structures exceed the original baseline is flat out staggering. Looking into the project details, one big factor was more intensive use and occupancy figures than modeled. This continues to point out the flaw of expecting too much from energy modeling that we have previously discussed. In addition, we continue to see the problem of trying to draw averages from lumping apples and oranges together. As someone recently said in a meeting on green building, drawing statistically across these projects appears far less educational and valid than more in-depth case study analysis of the specific individual projects.
(Hat tip to our friend Heidi Schwartz for commenting on this and sending us the report)