Determinants of Lighting Quality II: Research and Recommendations

Lighting practitioners and laypeople alike believe that the quality of the luminous environment can influence task performance, comfort, and well-being. There is mounting concern that the quality of the lit environment will decline in parallel to lighting energy use as energy codes and standards come into effect; however, there is no consensus about what constitutes good lighting quality. Broad agreement exists that illuminance, luminance, luminance distribution, uniformity, glare control, flicker rate, and spectral power distribution are the important dimensions of the luminous environment. This literature review relates these variables to behavioural outcomes such as visual comfort, task performance, preferences, and well-being. Comparison of these studies to the recommended practices for lighting design in offices reveals little agreement. Psychologists and other behaviour scientists, although not generally accustomed to participating in the writing of codes and standards, have an important role to play the development of these regulations, in ensuring that the best possible knowledge base is applied to the establishment of working and living conditions to support people. Résumé Tout comme les profanes, les praticiens de l’éclairage croient que la qualité de l’environnement lumineux peut influer sur l’accomplissement des tâches ainsi que sur le confort et le bien-être des personnes. On s’inquiète de plus en plus de la diminution prévue de la qualité du milieu éclairé à la suite de la mise en place de codes et normes réglementant l’utilisation de l’énergie; cependant, il n’y a pas de consensus sur ce qui constitue une bonne qualité d’éclairage. On admet généralement que l’éclairement, la luminance, la répartition de la luminance, l’uniformité, le contrôle de l’éblouissement, la vitesse de papillotement et la répartition de la puissance spectrale sont les dimensions importantes de l’environnement lumineux. Dans cette étude de synthèse, ces variables sont rattachées aux effets de l’éclairage sur le comportement, par exemple le confort visuel, l’accomplissement des tâches, les préférences ou le bien-être. La comparaison de ces études avec les façons de faire préconisées au niveau de la conception de l’éclairage des bureaux montre qu’il n’existe guère d’accord à ce sujet. Même s’ils n’ont généralement pas l’habitude de participer à la rédaction des codes et normes, les psychologues et autres spécialistes du comportement ont un rôle important à jouer dans l’élaboration de cette réglementation; en effet, ils doivent veiller à ce que la meilleure base de connaissances possible soit mise en place en vue de l’amélioration des conditions de travail et de vie. Lighting Research & Recommendations page 3 Determinants of Lighting Quality II: Research and Recommendations Surveys of office employees consistently report that lighting is among the more important features of office design and furnishings (e.g., “Office lighting”, 1980; Spreckelmeyer, 1993). Likewise, in the professional community of lighting designers and illuminating engineers, there is a long history of speculation that the quality of the luminous environment can influence task performance, comfort, and well-being (e.g., Miller, 1994; Wagner, 1985). Electric lighting consumes up to half of the electricity consumed in commercial buildings (Eley, Tolen, Benya, Rubinstein, & Verderber, 1993), but this percentage will decline as energy codes come into effect (e.g., American Society of Heating, Refrigeration, and Air-Conditioning Engineers [ASHRAE], 1989; Canadian Codes Centre, 1995). These codes regulate energy consumption in new buildings and also influence current practice in renovations and retrofits. Before the 1970s energy crisis and the introduction of computers into workplaces (which require less ambient illumination than paperwork), lighting power densities of 30 W/m (2.8 W/ft) were typical in North America. Today, values of 17 W/m (1.6 W/ft) are usual, and lower levels are easily attained with commonly-available equipment. The first strategy used to reduce lighting energy consumption in th early 1970s was delamping. In this practice, one fluorescent lamp was removed from a 2-lamp fixture; in some cases, alternate luminaires in a row were disabled. This reduced overall light levels (illuminance) and produced uneven distributions of light (luminance distribution). Energy was saved, but lighting quality declined. Lighting designers of the day decried this simplistic approach to conservation (e.g., Benya & Webster, 1977; Chase, 1977; Florence, 1976), and concern about the quality of the lit environment in the context of energy conservation persists today. Despite ongoing discussion throughout the 1980s ("Illumination Roundtable III", 1984), defining and debating lighting quality remains a contentious issue among the lighting community. Panel discussions, workshops, and seminars occur at nearly every major conference. The fear that lighting quality will decline in parallel to lighting energy use is difficult to address because there is no consensus about what constitutes good lighting quality. Some believe that no agreement on this issue is possible (e.g., Erhardt, 1994). One survey found only moderate consensus among lighting practitioners about the quality of computer-simulated lighting designs; differences between cultural groups were also observed (Veitch & Newsham, 1996a). There is, however, broad agreement about the important dimensions of the luminous environment. These are illuminance, luminance, luminance distribution, uniformity, flicker rate, and spectral power distribution (cf., CIBSE, 1994; NUTEK, 1994; Rea, 1993). Lighting system characteristics such as individual control, indirect versus direct lighting, and the use of daylight are also thought to contribute to good-quality lighting. Table 1 summarises three recommended practice documents for office lighting, one each from the United Kingdom, Sweden, and North America. These recommendations are based on consensus among committee members, and are notorious for their weak link to published research (Boyce, 1987). In a companion paper (Veitch & Newsham, 1996b), we argue that existing attempts to model or to predict lighting quality from luminous conditions have failed because of poor science: lighting research has tended to use abstract tasks for visibility measurements, a narrow range of behavioural outcomes, and inadequate specification of the population to which the data apply. We propose a behaviourally-based definition of lighting quality, in which lighting quality Lighting Research & Recommendations page 4 is defined as the degree to which the luminous environment supports the following requirements of the people who will use the space: • visual performance; • post-visual performance (task performance and behavioural effects other than vision); • social interaction and communication; • mood state (happiness, alertness, satisfaction, preference); • health and safety; • aesthetic judgements (assessments of the appearance of the space or the lighting). In this paper, we review the empirical evidence that relates these outcomes to the important dimensions in the luminous environment listed above, in an attempt to define the conditions that are associated with good lighting quality. Economic considerations have driven much lighting research, with the result that the vast majority of investigations have considered lighting for offices, with relatively few investigations occurring in other settings. Accordingly, this review focuses on office lighting applications, although studies from other settings are included where their results are relevant. Research on the Luminous Environment Luminance Luminance is the quantity of luminous energy propagated in a given direction by a point on a surface. Colloquially, this is what is generally meant when we speak of the brightness of an object, although this use confuses the photometric quantity and the sensation of brightness, which depends on the state of adaptation of the eye as well as the luminance of the object (Rea, 1993). The visual system adapts to changes in the ambient luminance by changes in pupil size and the responsivity of retinal photoreceptors. For example, pupil size decreases sharply upon going outside at high noon on a sunny day. The long history of lighting research is dominated by investigations of the relationship between luminance and visual performance (e.g., Blackwell, 1959; Boyce, 1973; Rea & Ouellette, 1991), with the result that we understand well how light levels affect visibility. It is well established that visibility relates to four variables: luminance, task/background contrast, task size, and the age of the observer. Colombo, Kirschbaum, and Raitelli (1987) suggested that a fifth variable, blur, be added to the visual performance model. Visibility measured using reaction times to a visual stimulus (Perry, Campbell, & Rothwell, 1987; Rea, Boyce, & Ouellette, 1987; Rea & Ouellette, 1988) and the time required to perform a number-checking task (Rea & Ouellette, 1991) have contributed to this understanding. Visibility is poor when the task luminance is low, but above a certain level of stimulation it quickly saturates. Perry et al. (1987) suggested that this change relates to a shift from rod to cone-based processing of retinal stimulation. The relationship between contrast and visibility is similarly asymptotic. For a given adaptation luminance there is a contrast value above which visibility is almost invariant; the drop-off below this value is sharp (Rea & Ouellette, 1991). The shape of this “plateau and escarpment” depends also on the size of the object being viewed, with smaller objects being most difficult to see and most adversely affected by reductions in luminance or contrast. 1 One study found that extremely high task contrast reduced visual performance on one task, which the authors suggested might be attributable to a dazzle effect for a spatially patterned task printed in intense black on white (Stone, Clarke, & Slater, 1980). Lighting Research & Recommendations page 5 As age clouds the lens, retinal illuminance declines; thus, the effective adaptation luminance is lower for older adults than younger ones. For this reason, older adults generally will require better contrast, higher task luminance, or larger objects to obtain the same visibility level as a younger person. This decrement in vision is noticeable around age 40 (Guth & McNelis, 1969). Luminous conditions that influence visibility may also affect other important variables. Rea, Ouellette, and Kennedy (1985) found that participants tended to modify their posture to maintain visual performance under luminous conditions that otherwise would reduce task visibility. This has important implications for offices and other workplaces and is worthy of more detailed examination. Awkward or slouching postures can lead to orthopaedic or other health problems that are painful and debilitating, and which are expensive for employers and society in terms of absenteeism, lost productivity, and health care costs.