Lighting, Performance, and Health
Typically, when addressing lighting requirements, guidelines for green schools focus on energy-efficient lighting systems and the use of daylighting to further conserve energy. They also encourage access to views through the installation of windows.
For purposes of learning, performance, and productivity, lighting in a school building should allow people to see to read, to see others with whom they are communicating, and to perform other visual tasks associated with learning, teaching, and school administration. Lighting can be provided by electric sources or by daylight through windows, clerestories, and skylights. Typically, school buildings use a combination of electric lighting and daylight.
Electric lighting systems are made up of a series of components: luminaires, lamps (incandescent, fluorescent, high-intensity discharge [HID]), ballasts (except when using incandescent lamps), and controls. Electric lighting systems differ in the amount of power they require to operate and in the amount and direction of light they are able to generate for the design objectives. They also vary in their initial cost, ease of maintenance and commissioning, and expected life. Also important, electric lighting systems vary in their ability to provide good color rendering, low glare, low flicker, and low noise.
Fluorescent lighting systems are the most prevalent sources of general illumination in schools. Modern fluorescent systems (T8 and T5 lamps with electronic ballasts) can provide low cost, long life, high efficacy, good color, low noise, and low flicker. Other sources of illumination, including incandescent and HID, can be specified in schools to best accomplish specific design objectives, from outdoor applications such as sports fields to illuminating pictures or works of art (Rea and Bullough, 2001).
Windows are an important part of the school design.6 First and foremost, windows provide a view to the outside. They also provide high light levels and, when properly located, ideal eye-task-lighting geometries for reflective visual tasks (i.e., those other than self-luminous displays, such as computer screens, and audio visual presentations). Skylights and clerestories can also be beneficial sources of illumination even though they do not provide a view.
A key difference between designs with electric light and/or daylight is that electric light is almost always static, whereas daylight is ever changing over the course of a day, with weather conditions, and with season. Daylight will also be different from one school to another, depending on building orientation and site, local climate, and latitude, so a “cookie cutter” building design will rarely provide ideal lighting conditions. The dynamic nature of daylight, together with its wide range of intensities and geometries, demands a dedicated understanding of its interactions with the building and its spaces. In some circumstances it may be desirable to conduct detailed lighting, heating, and cooling simulations in order to gain such an understanding. The potential benefits of higher lighting levels, excellent color
and form rendition, views, and the temporal lighting qualities may justify this careful design requirement and additional analyses.
LIGHTING AND ITS IMPACT ON THE VISUAL AND CIRCADIAN SYSTEMS
When evaluating the performance of any lighting system, electric or daylight, its impact on two biological systems—the visual and the circadian—needs to be considered together with the physical attributes of light that differentially affect these systems (Figure 4.1).
The Visual System
The visual system functions as a very quick remote-sensing mechanism that alerts us to changes in the environment and enables us to identify threats and opportunities around us. The visual system is fairly well understood for normal adult populations in regard to the effects of light on both appearance (what things look like) and visual performance (how well
visual information is processed). For example, there is a complete model of visual performance available for predicting the impact of background luminance (light level), target contrast, target size, and observer age from 18 to 65 years (Rea and Ouellette, 1991). Presumably, most school-age students should behave like 18-year-olds in regard to visual performance, but this has not been systematically studied. In general, given the characteristics of the visual response functions, it can probably be concluded that most lighting and task conditions are adequate for normal student visual performance. Going the next step, however, is more tenuous because there is no evidence that the quality or quantity of light directly affects student learning performance (Larson, 1965; Demos et al., 1967; Boyce et al., 2003).
Lighting for Visual Performance
The performance of any lighting design cannot be effectively evaluated solely on the basis of the source of illumination or the individual components needed to create the entire lighting system. Instead, lighting performance should be evaluated within the integrated system of enclosure design and controls, space geometry and finish, and fixture components in relation to the task requirements.
For a large majority of the people working in buildings, lighting for vision during the day is quite adequate, in part because people have very flexible visual systems and adjust their posture in response to the available lighting conditions. For example, the dimmer the light, the closer one holds the reading materials to maintain a constant ability to read. Experiments in the laboratory show that people will systematically adjust the eye-to-task geometry to maintain good task visibility, either by moving closer to the visual task or shifting posture to avoid reflected glare (Rea et al., 1985). A flexible visual system combined with a flexible body provides most people with the ability to adapt to less than ideal lighting environments.
A significant minority of the school-age population may not have properly corrected eyesight, however. Students without properly corrected eyesight may not be able to take full advantage of adaptive strategies to see learning materials in the classroom. In a large study using the 1996-1997 National Health Interview Survey, Kemper et al. (2004) determined that approximately 25 percent of school-age children in the United States have corrective lenses. They showed that the prevalence of corrective lenses is related to several population factors such as age (older children are more likely to wear corrective lenses), ethnicity (black and Hispanic children are less likely to wear corrective lenses), income (poorer children are less likely to wear corrective lenses), and gender (girls are more likely to wear corrective lenses than boys). Of particular note, insurance coverage appears to be a major factor in the prevalence of corrective lenses in school-age children. The actual percentage of corrective lenses probably underestimates the number of children who need them. Moreover, it is unknown whether students’ corrective lenses actually have the proper refraction. Because lighting and task variables (e.g., particularly task target contrast and size) as well as proper refraction determine visual performance, it seems likely that at least some students are not able to adequately see learning materials in the classroom.
Within the context of the school environment, then, adequate lighting for the majority of the students may be insufficient for some fraction of school-age children. It could be hypothesized that for those students who do not have the proper refraction, daylight may
offer a significant advantage by providing higher light levels and better geometries than would otherwise be present from electric lighting alone. However, the potential advantage of daylight in classrooms for improving visual performance of children with or without properly corrected eyesight has not been systematically studied.
The adult (teaching and administrative) population will also likely have problems with refractive errors. Normal aging involves a continuous loss of visual accommodative ability, known as presbyopia, from about 20 years of age to approximately 65 years (Weale, 1992). Until about age 45 the gradual loss in ability to focus on near objects is hardly noticed, but after this approximate age nearly everyone begins to adopt new strategies to see small targets. Instead of getting closer to an object to see it as they did when they were younger, people with presbyopia actually move the object further away from their eyes, or they place the object under a bright light, usually provided by a window or skylight. Eventually, everyone seeks optical aids, such as bifocals or reading glasses, to see normal print, but use of bright light from the right direction will continue to be a strategy employed by older people to see small targets throughout their lives.
Glare and Visual Performance
In regard to glare, there is much less certainty in predicting visual comfort, even in adults, than in predicting visual performance (Rea, 2000; Boyce, 2003). A clear distinction is made between glare that reduces visual performance (disability glare) and glare that does not (discomfort glare). Disability glare can be precisely predicted for a given individual and for the general population. As would be expected, disability glare becomes more problematic with age owing to changes in the optical media of the eye, particularly from scattering of light within the eye by the crystalline lens (Weale, 1963).
Windows are the largest sources of glare in a classroom. However, glare can be controlled by using fixed overhangs in conjunction with blinds or window treatments that can be manually operated. Methods to control light from skylights and clerestories are also needed because of ever-changing lighting geometries and light levels as the day progresses.
Formulas do exist for calculating discomfort glare, and these are often used to characterize the lighting layout for a space using commonly available lighting software. However, collective understanding of the mechanisms underlying discomfort glare is rather poor (Boyce, 2003). It appears that psychological phenomena contribute significantly to discomfort glare so that, for example, bright flashing lights in an office are highly uncomfortable, but the same lights can be highly desirable in a night club for dancing. Therefore, although disability glare is the same in both applications, discomfort glare is not. In this context, and given the strong psychological component to discomfort glare, predicting visual comfort in school-age children is an area that requires more research.
DAYLIGHTING AND STUDENT LEARNING
Several well-designed studies investigating the effect of daylighting on student performance were conducted by the Heschong-Mahone Group between 1999 and 2003. In the 1999 study, data were obtained from three elementary school districts located in Orange
County, California; Seattle, Washington; and Fort Collins, Colorado (Heschong-Mahone, 1999). The study looked for a correlation between the amount of daylight provided by each student’s classroom environment and test scores. Test results for more than 21,000 students in these districts were analyzed. Demographic data sets, architectural plans, aerial photographs, the presence of skylights, maintenance records, and daylighting conditions for more than 2,000 classrooms were among the factors reviewed.
The study developed a regression model for approximately 150 independent variables (e.g., teacher salaries, grade level, attendance), including available daylight represented by five different levels, or “daylight codes.” Although the regression analysis leads to a prediction that an increase in the value of the daylight code can increase scores in both math and English by more than 20 percent, a closer examination of the results shows that only 0.3 percent of the variance in the regression model is explained by daylight code (Boyce, 2004). This is a very small effect and one that cannot be justified as reliable. As noted below, these results could not be replicated in a subsequent study.
A reanalysis of the Capistrano and the Seattle school districts’ data was undertaken in 2001 to look at additional variables that might have a confounding influence, including teacher assignments (Heschong-Mahone, 2001). In 2003 a third study was undertaken to see whether the original methodology and findings would hold when data came from a school district with a different climate and curriculum. The Fresno California school district was used. The preliminary statistical analyses replicated the structures of the models used in previous studies. In the Fresno study, the holistic variable called the Daylight Code “was not significant in predicting student performance. It had the least explanatory power of the variables considered, and the lowest significance level” (Heschong-Mahone, 2003, p. viii).
The authors proceeded with more detailed multilinear regression (statistical) analysis to see whether they could gain some insight into why the Daylight Code was not significant in Fresno as it had been in earlier studies. Among the authors’ conclusions were that sources of glare negatively affect student learning; direct sun penetration into classrooms, especially through unshaded east or south facing windows, is associated with negative student performance, likely causing both glare and thermal discomfort; blinds or curtains allow teachers to control the intermittent sources of glare or visual distraction through their windows; when teachers do not have control of their windows, student performance is negatively affected (Heschong-Mahone, 2003, p. ix). They summarized that
Characteristics describing windows were generally quite stable in their association with better or worse student performance. Variables describing a better view out of windows always entered the equations as positive and highly significant, while variables describing glare, sun penetration, and lack of visual control always entered the models as negative. (Heschong-Mahone, 2003, p. viii)
Because of the inconsistent results of this limited number of well-designed studies, there is insufficient evidence at this time to determine whether or not an association exists between daylighting and student learning.
LIGHTING AND THE CIRCADIAN SYSTEM
The circadian system involves biological rhythms that repeat at approximately 24-hour intervals. The behavior of all terrestrial species, including humans, is driven by an internal clock synchronized to the solar light-dark cycle. Indeed, light is the primary stimulus to the internal clock. The circadian system regulates not only overt daily patterns of behavior such as activity and rest, but also our bodies at the cellular level, regulating functions such as the cell cycle (Moore, 1997).
Current lighting technologies and lighting standards are designed exclusively for providing visual sensation. However, light affects the visual system very differently than it affects the circadian system. Relative to the visual system that underlies conventional photometry and all lighting standards, the circadian system needs a much higher light level on the retina for activation (McIntyre et al., 1989a,b); it has a peak spectral sensitivity to much shorter wavelengths (Brainard et al., 2001; Thapan et al., 2001); it has greater sensitivity to light in the inferior retina (viewing the sky) than in the superior retina (Glickman et al., 2003); it requires much longer exposures for activation (McIntyre et al., 1989a,b; Rea et al., 2002); and, most important, it is differentially sensitive to light depending on the time of day (Jewett et al., 1997).
There is a growing body of literature indicating that the effect of light on circadian regulation can affect productivity as well as health. Seasonal affective disorder (SAD), or the “winter blues,” is recognized by the medical community as a psychiatric disorder. Apparently, seasonal reductions in the amount of daylight available in the winter at extreme northern (and southern) latitudes can induce depression (Rosen et al., 1990). Light treatment, typically provided with bright light from electric lighting systems, is recognized by the medical community as the preferred method of treating SAD (Rosenthal et al., 1985).
The incidence of SAD, or winter blues, in school-age children is poorly documented, although it has been reported that adults who experience SAD also experienced it as a child. It seems too that postpubescent young women are more likely to experience SAD (Rosenthal, 1998). Depending upon latitude, between 4 percent and 30 percent of the adult population, usually women, experience some symptoms of seasonal depression (Rosenthal, 1998), which in turn, might affect teachers. Less learning might be expected from those children who experience symptoms of seasonal depression, so lighting may play a very important role in the design of a green school at northern latitudes during the winter months. Systematic attempts to alleviate seasonal depression in children through lighting design have not been undertaken, but these early findings suggest a reconsideration of the role that light, particularly daylight, plays in the classroom.
Nearly half the population experiences some form of sleep disorder (National Sleep Foundation, 2005). Poor sleep directly affects a person’s ability to perform tasks and learn new tasks (Jennings et al., 2003; Heuer et al., 2004). Light and dark have a dramatic impact on sleep quality (Turek and Zee, 1999; Reid and Zee, 2004). Adolescents in particular commonly go to sleep late (after midnight) and have difficulty getting up early (before 7:00 a.m.) to go to school (Carskadon et al., 1998). In extreme cases this difficulty in falling asleep early and getting up early is diagnosed as delayed sleep phase syndrome (DSPS). Many students with DSPS must get special training or even repeat grades because of poor attendance and poor performance. Light is a recognized treatment for this disorder, and a
regular light-dark cycle may have broader implications for sleep quality in a larger group of children.
Recent research at the other end of the age spectrum shows that light treatment can consolidate sleep and increase sleep efficiency during the night in older people (Satlin et al., 1992; Fetveit et al., 2003; van Someren et al., 1997; Figueiro and Rea, 2005). A regimen of bright light at school during the day, together with dark nights at home, may increase student attendance and performance. Here again, however, there have been no systematic studies in school-age children.
DAYLIGHTING, VIEW, PERFORMANCE, AND HEALTH
Studies have been conducted using subjective ratings from adults providing evidence that people like views from windows (Markus, 1967; Jackson and Holmes, 1973; Ne’eman and Longmore, 1973; Collins, 1975; Ludlow, 1976; Cuttle, 1983; Heerwagen and Heerwagen, 1986; Leslie and Hartleb, 1990, 1991; Boubekri et al., 1991) and that real estate prices are higher for architectural spaces with a view from windows (Boyce et al., 2003). Kuller and Lindsten (1992) studied children’s health and behavior in classrooms with and without windows for an entire academic year. At the end of the study, they concluded that work in classrooms without windows affected the basic pattern of the hormone cortisol, which is associated with stress. The authors concluded that windowless classrooms could have a negative effect on children’s health and concentration. This finding is strictly suggestive, however, because no direct relationship between cortisol levels and student performance and health was established (Rusak et al., 1997).
FINDINGS AND RECOMMENDATIONS
Finding 4: In regard to lighting, performance, and health, the committee has found the following:
Daylight is a special light source because it may provide a view (through a window), high light levels, and good color rendering, and it is ever-changing. Direct and reflected sunlight can create significant visual problems if windows, skylights, and clerestories allow in too much or too little light. However, such problems can be controlled with manual blinds and other types of window treatments.
There is good evidence from studies of adult populations that the visual conditions in schools resulting from both electric lighting and daylighting are adequate for most children and adults.
There is, however, concern that a significant percentage of students in classrooms do not have properly corrected eyesight, and thus, the general lighting conditions suitable for visual functioning by the average student may be inadequate for those students without properly corrected eyesight. It could be hypothesized that daylight may benefit these children by providing higher light levels and better geometries than would otherwise be present from electric lighting alone.
However, the potential advantages of daylight in classrooms for improving visual performance of children with or without properly corrected eyesight has not been systematically studied.
Because of inconsistent results and the small number of well-designed studies, there is insufficient evidence at this time to determine whether or not an association exists between daylighting and student performance.
A growing body of evidence suggests that lighting may play an important nonvisual role in human health and well-being through the circadian system. However, the effect of light on health through circadian regulation of sleep, depression, and cell cycle has not been directly studied in children.
Recommendation 2: To determine the potential and actual performance of a lighting system, the entire system should be assessed because the total performance cannot be effectively evaluated based solely on the source of illumination or on the individual components needed to create the entire lighting system.
Recommendation 3: For green schools in which the lighting strategy is to use daylight extensively, control systems that can be easily operated, such as manual blinds or other types of window treatments, should be specified in order to control excessive sunlight or glare.