Evidence-Based Design: Natural Light, Student Performance and Health

Lisa Heschong is a principal of Heschong Mahone Group and a licensed architect who has divided her professional practice between energy research, writing and building design. As a researcher, she led the project teams that analyzed the impact of daylighting on human performance for the Daylighting and Productivity Studies, funded through PG&E and CEC. She also led the team that analyzed baseline lighting characteristics and created a computer model of lighting energy use for the state of California Energy Commission. She is the author of Thermal Delight in Architecture, a co-author of the Advanced Lighting Guidelines, the CHPS Best Practices Manual, and the Skylighting Guidelines, all Web-based publications. As a lighting expert, she has developed the successful web-based training program for the Federal Energy Management Program (FEMP) and conducted workshops across the country for DOE. She has published scholarly papers and conducted numerous lectures and workshops across the country on issues of daylighting, high performance design, energy efficiency and human comfort. As an architect, Ms. Heschong has managed projects to design high-rise office buildings, K-12 schools and residences. She has consulted on numerous school designs through the Bright Schools program of the California Energy Commission, and reviewed school proposals for the Office of Public School Construction. She also taught studio design at the Architecture Department of the University of California at Berkeley. She is experienced working with multi-disciplinary building design teams. She understands the construction process and the dynamics of getting a building project initiated, funded, designed and occupied. Ms. Heschong is a member of IESNA, and is Lighting Certified. She was awarded her B.Sc. at UC Berkeley, Summa Cum Laude, and her Master of Architecture degree at the Massachusetts Institute of Technology with the AIA Medal.

"I have spent a major chunk of my career, both at the beginning and now, the last decade-and-a-half, involved in research and consulting. Our company specializes in building energy efficiency, but as an architect, my interests are even broader than just energy efficiency. I care enormously about occupant health and public health. One of the unique things we have done as a company is evaluate the success of various utility and government programs to promote innovation in the building industries. As a result we’ve learned how to look at buildings as populations, or as a social infrastructure rather than as a particular design. So our consulting is not consulting on the design of a particular building so much as it is looking at how policy and design trends are influencing this infrastructure. It is rather a different niche than most architects find themselves in."

Hava Leisner : The landmark “Daylighting in Schools Study” from 1999 researched the correlation between natural light diffusion in the classroom and the impact upon student performance. Was that the study that showed the significant increase in student performance?

 Lisa Heschong: That was the study. Our motivation was to see if we could find a quantifiable connection between daylight and human performance. We studied schools because schools have a ready-made way to test student performance. The physical environments are so similar — practically identical small classrooms — that we could control for this one variation that we wanted to study: daylighting. We didn’t know, obviously, if we would find any results, but we also didn’t know how to describe daylighting. What was important about it? Was it the view? Was it the sense of having changeability in the space? Was it the amount of illumination? Was it the quality of the illumination? Was it the spectrum of the illumination? I mean ‘daylight’ is an easy thing to say, but we didn’t really know the particulars about daylight and what its mechanism might be.

For PG&E, we set out to find schools to study that included daylighting both by windows [or sidelighting] and by skylights, or toplighting. And we reasoned that if we got similar results from classrooms that had both the windows and the skylights, or the toplighting, it was probably a daylighting affect; it was probably something having to do with the quality of the light coming from a natural source. Whereas if we got a different result between the affect of windows or the affect of skylights, if one was positive and one was negative, then we would be able to say ‘well maybe it’s the view, or maybe it’s the intensity of the light,’ or we could start distinguishing some of those characteristics.

So, in this first study, the 1999 study, we selected three districts that had both conditions [Capistrano USD in California ; Poudre R-1 SD in Fort Collins , Colo. , and Seattle , Wash. , Public Schools]. We got a positive association: more daylight from skylights, or toplighting, better student results; bigger windows, more daylighting from windows, better student results. And so our assumption was it’s probably the daylight — something about the characteristics of natural light coming from the sky — that is having these positive affects.

There is a middle study in there, which is the re-analysis report. When we got the findings from the first study, PG&E recognized that this was groundbreaking work and they got very cautious and asked us to go through another level of peer review, which was a peer review process led by statisticians and folks at Lawrence Berkeley National Labs to basically throw rocks at the study and see if there were things that we left out. We answered all their questions. The biggest challenge that we could not answer was were better teachers showing up in the more daylit classrooms and is that why kids were doing better. So we went back and collected information about the teachers at Capistrano and then we re-ran the equations and we did even more sensitive statistical tests on the equations.

The answer came back, ‘no, it wasn’t the teachers.’ The teachers were not being sorted by better teachers in more daylit classrooms, and they weren’t contributing to this affect that we were finding. This affect stood. It was the kind of definitive piece that said we’ve answered all the challenges from the peer review and found a central tendency of 21 percent improvement in student performance of those students in the classroom with the most daylighting compared to no daylighting.

Then the California Energy Commission funded us to do a third study and we worked in yet another district, a fourth district, where we could use the same methodology and see if the results still held. That district had no top-lit classrooms, no skylights and no clerestories.  

HL: And that was intentional?

LH: No it wasn’t. We chose Fresno, Calif., because it is a very large district and they had the right kinds of testing protocols for our methodology. They used the same testing company and testing protocols as Capistrano, so that would make our outcome matrix more comparable. And there seemed to be a sufficient range of daylighting conditions.

We did a preliminary survey to see what the classrooms looked like and what the range and conditions would be. When we finally filtered down to our final study population, we lost a lot of that diversity, unfortunately. That was just a function of which students we could study at Fresno . We had a much smaller number of aggressively daylit classrooms, so we didn’t have as broad a range, you know, nice distribution of conditions. We had a distribution that was the opposite of a bell curve — with a greater number of daylit classrooms and non-daylit classrooms and fewer classrooms in between.

When we went out to Fresno , we collected a lot more information about the classrooms, so we could describe the windows in much more detail. We knew the size, the length, the height, the tint, the quality of the view through the window, and the illumination levels inside the classroom. We also collected information about the lighting systems, the noise caused by the lighting systems, the noise caused by the HVAC system, whether the teachers had control of the fans, the windows, the lights, and so on: a massive amount of information about the physical characteristics of the classrooms. We had 450 classrooms in the study and we had something like 75 descriptors of those classrooms.

 HL: So those would be like variables?

 LH: Yes. So for instance, we knew if the teacher turned a fan on and off, or if it was done automatically. We knew if there was a sink, a telephone, or a T.V. in the classroom. We knew the percentage of carpet on the floor, and if it was fully carpeted or just a little area rug. Was it 10 percent carpeted or 100 percent? And then we used our same methodology.

First, we controlled for all of the demographic variables of the students, then we controlled for all of the educational and demographic information about the teachers. Their education level, and so on. Then we start adding information about the classrooms to see what about the physical characteristics of the classroom were associated with better or worse student performance. Lots of information came back. That’s what those long tables are in the report. We learned a lot more than just about daylighting.

HL: Even though there was a much smaller number of daylit classrooms, considering all the variables you collected, were they able to make your study, in your eyes, equally robust because so much was accounted for?

 LH: We didn’t learn as much about daylighting per se as we would have liked. But we did learn a lot about positive and negative characteristics of windows and other aspects of the physical conditions of classrooms.

In terms of windows, the message that came across very clearly was that windows with a better view, that allow teachers to control for glare and distractions with blinds or curtains, and that prevent glare and sun penetration, because they are well shaded or they don’t look out at glary things, are strongly associated with better student performance. So we had a variety of descriptions that were positive: larger view windows, view of vegetation, far vs. near views, view of human activity.

Anything associated with glare, such as sun penetration or lack of blinds on the windows, are associated with negative student performance.

HL: I know in the landmark 1999 study you looked at operable windows vs. fixed windows, and that progress rate. Did you do that at Fresno as well?

LH: We did. And in general Fresno students seemed to benefit from continuously mechanically ventilated classrooms. All the classrooms in Fresno are air-conditioned. And as you probably know, Fresno does not have very good indoor air quality and it also has a very high-rate of asthma. It is sort of a hot spot for asthma.

HL: Is it dusty out there?

LH: Very dusty. So what we found is that operable windows tended to be associated with negative performance. Ventilation systems that could be turned off by the teacher tended to be associated with negative performance. Schools that had automatic continuous ventilation were doing better in Fresno. And one of the other things that is very interesting about the study is that we had the opportunity to go back and observe these schools and classrooms in operation to see how the teachers were running the classroom and how they were teaching in the classroom.

We sampled 40 classrooms and observed them for a few hours. What we found was an enormous challenge for the teachers. They are sort of between a rock and a hard place. They would have a number of systems and they could only operate one or two of them well. In other words, they would have to choose to ventilate their room but accept horrible acoustic conditions or they would have to choose to have a quiet room but no ventilation. They could choose to have their windows open for ventilation but bad playground noise and cold rooms. This is in Fresno, and I don’t think that it is that different from other districts in California. Teachers would have to choose to ventilate their room but accept horrible acoustic conditions or they would have to choose to have a quiet room but no ventilation.

A lot of these teachers were forced into making a few bad decisions in order to optimize something else in their classroom. One of the things that we found in Fresno was that the classrooms that tended to have the best daylighting conditions also tended to have some of the worst acoustic conditions. And the classrooms that tended to have the best acoustic conditions tended to have the worst daylighting conditions.

HL: That’s huge.

LH: Yes. So for example, you would have a classroom from the ’60s that was designed with windows on both sides and a big volume space and nice ventilation going through but lots of hard surfaces, which made it very resonant. Then they would have a bilingual aid in the back of the classroom translating simultaneously while the teacher is talking. And so the teacher’s voice and the aide’s voice are basically causing interference for each other in this very resonant space.

What we observed happen was in the old-style schools the bilingual aide was most often located directly in the classroom. So there were at least two activities going on in the classroom simultaneously. In the newer schools that were often designed on the open plan or with shared workrooms or pullout spaces, the bilingual aides would work outside of the classroom so they didn’t have acoustic conflict. That’s something you can only really see by going and observing how the classrooms are being operated.

School Construction News will continue its dialogue with Ms. Heschong as the Q&A in the Nov/Dec 2004 issue, where she discusses rising energy costs and what those in the school industry can do to conserve resources and combat costs.