|
This study reports on an investigation into
the influences indoor physical environment has on office worker
performance. It is particularly concerned with the potential
contributions of windows and daylight to improved performance by office
workers. Two different studies were conducted at the same
organization, the Sacramento Municipal Utility District. The first
study looked at 100 workers in an incoming call center, whose performance
was continuously tracked by a computer system and measured in terms of time
to handle each call. The second study examined the performance of 200 other
office workers on a series of short cognitive assessment tests, taken at
each individual's desktop computer.
The study sites provided a range of daylight,
view and ventilation conditions, while providing a relatively uniform
environment for other potential influences on worker performance. All of the
office work considered was computer-based, based on self-illuminated tasks.
Extensive data was collected about the physical environment at each office
worker's cubicle. Multivariate regression analysis was used to control for
other potential influences, such as age or employment status. A variety of
statistical models were tested to determine if any of the variations in
environmental conditions, either between workers or during different time
periods for a given worker, were significantly associated with differences
in worker performance.
The studies found several physical conditions
that were significantly associated (p<0.10) with worker performance, when
controlling for other influences. Having a better view out of a
window, gauged primarily by the size of the view and secondarily by greater
vegetation content, was most consistently associated with better worker
performance in six out of eight outcomes considered. Workers in the Call
Center were found to process calls 6% to 12% faster when they had the best
possible view versus those with no view. Office workers were found to
perform 10% to 25% better on tests of mental function and memory recall when
they had the best possible view versus those with no view. Furthermore,
office worker self reports of better health conditions were strongly
associated with better views. Those workers in the Desktop study with the
best views were the least likely to report negative health symptoms. Reports
of increased fatigue were most strongly associated with a lack of view.
Other variables related to view were also
found significant. In the Call Center higher cubicle partitions were
associated with slower performance. In the Desktop study glare
potential from windows was found to have a significant negative effect on
performance in three of the five mental function assessment tests. In the
three tests, the greater the glare potential from primary view windows, the
worse the office worker performance, decreasing by 15% to 21%, all other
things being equal.
Horizontal daylight illumination levels were
found to have an inconsistent relationship to performance, significant in
two out of eight metrics tested. Higher levels of daylight illumination were
found positive for Digit Span Backwards, a test measuring attention span and
short term memory, and negative when compared to changes in daily average
speed of handling calls for one of two study periods. The natural log of
daylight illumination levels was found to have the best mathematical fit to
the data, implying more sensitivity to changes at lower levels of
illumination and progressively less sensitivity at higher levels.
Ventilation status and air temperature were
also found to have significant, if intertwined and occasionally
contradictory, associations with worker performance. When variation in
hourly performance at the Call Center was considered, higher rates of
outside air delivery were significantly associated with faster handling of
calls.
Overall these potential influences on worker
performance were found to have high statistical significance in the models
tested. They are related to performance that is 1% to 20% better or worse
than average. All together information about the physical conditions
of the workers was able to explain about 2% to 5% of the total variation
observed in a measure of worker productivity (Call Center study) or in
performance on short cognitive assessment tests that were thought to be
related to worker productivity (Desktop study).
Even small improvements in worker
productivity are of great practical importance, and explaining 2%-5% of
total variation is not trivial. By way of comparison, all other available
information typically believed to predict performance such as demographic
characteristics or employment status was able to explain about 6% to 19% of
the variation in their performance. Thus the characteristics of the physical
environment represent about 1/8th to 1/3rd of our entire ability to predict
variation in individual worker performance.
Furthermore, changes in the physical design
of a space that may influence worker performance are likely to have great
persistence, continuing for the life of the building. When compared with the
costs, persistence and the certainty of other methods of increasing
productivity, constructing well-designed buildings may be attractively
cost-effective. As demonstrated in the study site, these same features can
also provide additional energy cost savings.
Both studies successfully measured variation
in office worker environmental conditions and related these to measured
office worker performance under actual employment conditions. The Desktop
study pioneered the use of computerized cognitive assessment tools to gauge
office worker performance in field conditions. The studies have shown that
indoor environmental conditions can have a measurable relationship to
changes in office worker performance and have established a range of likely
effect sizes that other researchers can use to refine the needs of future
studies. Other studies will be required to test if these findings can
be replicated in other settings and to explore potential causal mechanisms
between the environmental conditions and worker performance.
This study investigates whether daylight and
other aspects of the indoor environment in elementary school student
classrooms have an effect on student learning, as measured by their
improvement on standardized math and reading tests over an academic
year. The study uses regression analysis to compare the performance of
over 8000 3rd through 6th grade students in 450 classrooms in the Fresno
Unified School District, located in California’s Central Valley.
Statistical models were used to examine the relationship between elementary
students’ test improvement and the presence of daylight in their
classrooms, while controlling for traditional education explanatory
variables, such as student and teacher demographic characteristics. Numerous
other physical attributes of the classroom were also investigated as
potential influences, including ventilation, indoor air quality, thermal
comfort, acoustics, electric lighting, quality of view out of windows, and
the type of classroom, such as open or traditional plan, or portable
classroom.
Previous Studies
This study is the third in a series of
studies looking at the relationship between daylighting and student
performance. The first, Daylighting in Schools,[1]
which was completed for Pacific Gas and Electric in 1999, examined school
districts in three states. In Seattle Washington and Fort Collins Colorado,
where end-of-year test scores were used as the outcome variable, students in
classrooms with the most daylighting were found to have 7% to 18% higher
scores than those with the least. In San Juan Capistrano, California,
where the study was able to examine the improvement between fall and spring
test scores, we found that students with the most daylighting in their
classrooms progressed 20% faster on math tests and 26% faster on reading
tests in one year than in those with the least.
A second study, the Daylighting in Schools
Reanalysis Report[2] completed for the California
Energy Commission in 2001 further investigated the results from the
Capistrano school district. We investigated whether better teachers were
being stationed in more daylit classrooms, and thereby inflating the
importance of the daylight variable. In that district, we found that there
was no assignment bias of better teachers to more daylit classrooms.
Furthermore, the addition of information about teacher characteristics to
the original student performance models did not reduce the significance or
magnitude of the daylight variables. Among twelve models considered in that
study we identified a central tendency of a 21% improvement in student
learning rates from those in classrooms with the least amount of daylight
compared to those with the most.
Fresno Study
This study’s primary goal was to
examine another school district, one with a different climate and curricula,
to see whether the original methodology and findings would hold. We
collected more information about the lighting and daylighting conditions in
the classrooms, to allow us to test which attributes of a daylit classroom
were more likely to contribute to a “daylight effect,” if any. We
also wished to understand how other aspects of the indoor environment
affected student performance and interacted with daylight. To accomplish
these goals, this study gathered detailed information about classroom
conditions, including lighting and daylighting, HVAC, ventilation, windows,
surface coverings, view, and indoor air quality. Whereas we had done
on-site surveys only a sample of classrooms for the previous studies, for
this study we went on-site to measure attributes in every classroom,
surveying a total of 500 classrooms in 36 schools.
The preliminary statistical analyses
replicated the structure of the models used in the previous studies. They
used a holistic variable called the Daylight Code to rate classrooms by the
amount of daylight available throughout the school year. In these
replication models, the Daylight Code was not significant in predicting
student performance for Fresno. It had the least explanatory power of the
variables considered, and lowest significance level. Thus, we could not
replicate the Capistrano findings based on a similar model structure. We
proceeded with more detailed statistical analysis to see if we could
identify specific influences of school or classroom design on student
performance, and perhaps gain some insight as to why the Daylight Code was
not significant in Fresno as it had been in Capistrano, Seattle and Fort
Collins.
We used multi-linear regression analysis to
test a wide variety of variables to see which provided the best explanation
of student performance. Of the variables describing the physical conditions
of classrooms and schools, 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.
In addition, attributes of classrooms
associated with acoustic conditions and air quality issues followed a
similar pattern. Those variables representing sources of internal noise,
such as a loud HVAC system or a loud ballast hum from the lighting system,
were consistently associated with negative student performance, while
increasing the amount of carpet (which reduces acoustic reverberance) in the
classroom was associated with better student performance in reading.
Variables related to indoor air quality showed that in Fresno automatically
controlled mechanical ventilation (No Teacher Control of Fan) was positive,
while visible water damage or a surveyor assessment of musty air in the
classroom was negative.
Summary of Study Findings
The findings of regression models in this
study support the general conclusions that:
- The visual environment is very
important for learning.
- An ample and pleasant view out of a
window, that includes vegetation or human activity and objects in the
far distance, supports better outcomes of student learning.
- Sources of glare negatively impact student
learning. This is especially true for math learning, where instruction
is often visually demonstrated on the front teaching wall. Per our
observations, when teachers have white marker boards, rather than black
or green chalk boards, they are more likely to use them and children
perform better in math.
- 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.
- The acoustic environment is also
very important for learning. Situations that compromise student
focus on the lessons at hand, such as reverberant spaces; annoying
equipment sounds, or excessive noise from outside the classroom, have
measurable negative effects on learning rates.
- Poor ventilation and indoor air quality
also appear to negatively affect student performance. However, in FUSD
these issues are almost hopelessly intertwined with thermal comfort,
outdoor air quality and acoustic conditions. Teachers often must choose
to improve one while making another aspect of the classroom worse.
- Physical characteristics of
classrooms are just as likely to affect student learning as many other
factors commonly given much more public policy attention. Variables
describing the physical conditions of classrooms, most notably the
window characteristics, were as significant and of equal or greater
magnitude as teacher characteristics, number of computers, or attendance
rates in predicting student performance.
Problems with Daylit
Classrooms
We tested each statistical
model with and without the Daylight Code. When we added the Daylight Code
the other variables remained essentially the same, but the Daylight Code
always came in as significant and negative, telling us that there was some
characteristic of classrooms sorted by the Daylight Code that was associated
with a negative effect. Examination of the performance of individual
classrooms, considering all of their window characteristics plus the
Daylight Code, showed that there were three types of classrooms in Fresno
that were performing particularly well in relationship to their daylight
characteristics—finger plan classrooms, grouped plan classrooms and
portables—as long as they had no glare or other undesirable window
characteristics. Thus, classrooms with both the highest and the lowest
Daylight Code were seen to support better student performance.
Many potential explanations
for the negative influence of the Daylight Code were considered, and we went
back on site to see if there were any systematic reasons why students in
classrooms with a higher Daylight Code would perform worse, or those in
classrooms with a low Daylight Code would perform better. In this second
phase of the study, detailed examination of a number of potential
confounding variables, including view-related distractions, glare, operable
windows, radiant thermal comfort, indoor air quality and acoustic
performance were considered. To better understand the results of the
regression analysis, we visited 40 classrooms while they were in operation
and surveyed 116 teachers about their assessment of and operation of their
classrooms.
Overall, the daylit
classrooms in Fresno had some consistent problems that might have degraded
student performance, and which we believe did not exist in the previous
districts studied. The most compelling of these were the acoustic problems
created in the daylit classrooms. We found the classrooms with high daylight
codes to have reverberation levels above current national recommendations,
while classrooms with low daylight codes typically met or exceeded those
recommendations. This reverberation problem tended to be aggravated by the
presence of teaching assistants who provide in-class tutorials for
individuals or small groups. In low Daylight Code classrooms these tutorials
were often held outside of the classroom in conveniently adjacent common
areas, while in the high Daylight Code classrooms they took place in the
back of the classroom, raising the background noise level and making the
teacher’s voice less intelligible.
In addition, we noted
teachers in classrooms with a high Daylight Code were more likely to teach
with their windows open, primarily to compensate for poor temperature
control and to improve ventilation. These open windows allowed in more noise
from the outside, exacerbated by crowded schools running on multiple lunch
and recess schedules. We noted from the various regression models that, on
the one hand, continuous mechanical ventilation seemed to improve student
performance, while on the other hand, a higher percentage of operable
windows were associated with lowered performance. We hypothesize that the
poor outdoor air quality in Fresno [3], combined with the epidemic of
asthma in school children, suggests the preferred use of mechanically
filtered air rather than natural ventilation in FUSD.
We also considered
whether the problems we detected with daylit classrooms could be rectified,
and calculated the value of potential energy savings if daylit classrooms
were operated to reduce reliance on electric lighting. Acoustic analysis of
the daylit classrooms showed that the reverberance problem could be
corrected with the use of more sound-absorbing surfaces, such as carpet and
high quality acoustic tile. The use of dual pane low-e glazing on the
windows could simultaneously improve both the acoustic conditions in the
classrooms and thermal comfort. Energy analysis showed substantial potential
savings (1.1 kwh/sf) for retrofitting existing FUSD daylit classrooms with
photocontrols.
California
could achieve an additional 3300 to 4800 megawatthours (0.6 to 0.9 kwh/sf)
of energy savings statewide for each year that all new school construction
included good daylighting design with photocontrols. This would accumulate
to 33,000 to 48,000 megawatthours per year savings after ten years.
The
Importance of School Design Choices
These findings suggest the
importance school planners should give to the architectural design of
schools. The statistical models repeatedly demonstrate that physical
condition of classrooms and schools are just as likely to affect student
learning as many other factors commonly given much more public policy
attention. Variables describing the physical conditions of classrooms, most
notably the window characteristics, were as significant and of equal or
greater magnitude as teacher characteristics, number of computers, or
attendance rates in predicting student performance. The partial R2 of the
different variable types is also very informative. The one variable which is
specific to the individual—their fall test score—predicts about 10% of
the variation in the gain from fall to spring. The demographic
variables, which describe generic groups to which the individual belongs,
predict performance with an order of magnitude less precise, or about 1%
each. The physical characteristics of the schools again drop another order
of magnitude in predictive power, each significant variable describing on
the order of 0.1% of the variation in student performance.
However, even though the
physical characteristics of classroom have a very minor potential influence
on the performance of a given individual, they will reliably affect hundreds
or thousands of students over the life of the building, typically 50 years.
Since the design of classrooms is entirely within the control of the school
district, much more so than student or teacher demographics, optimized
design of schools should be a central concern for all new school
construction.
This study presents evidence that a major
retailer is experiencing higher sales in daylit stores than in similar
non-daylit stores. Statistical models were used to examine the
relationship between average monthly sales levels and the presence of
daylight in the stores, while simultaneously controlling for more
traditional explanatory variables such as size and age of the store, amount
of parking, local neighborhood demographics, number of competitors, and
other store characteristics. The retailer, who will remain anonymous,
allowed us to study 73 store locations in California from 1999 to 2001. Of
these, 24 stores had a significant amount of daylight illumination, provided
primarily by diffusing skylights.
This study was performed as a follow-on to a
similar study completed for Pacific Gas and Electric in 1999[4],
which found that for a certain retail chain, all other things being equal,
stores with skylights experienced 40% higher sales than those without
skylights. This study, on behalf of the California Energy Commission,
examined a second retail chain, in an entirely different retail sector, to
see if the original findings would hold in a new situation, and if we could
learn more about any daylight effect that might exist.
As a first step in this process, a simple
model with daylight as a yes/no variable, and using basically the same
format and inputs as the previous study, did not find a significant
correlation between the presence of daylight, and increased sales. We then
pursued the study in greater detail, adding more information to the model
and describing daylight on a continuous scale by the number of daylit hours
per year in each store.
The retailer in this study had a less
aggressive daylighting design strategy and also more variation in both the
range of daylight conditions and the range of store designs than the
retailer in the first study. For this study, we collected much more
detailed information about the characteristics of each store, and verified
all information on site. Neighborhood demographics and retail competition
were described using detailed, site-specific GIS analysis. Store
managers were interviewed and employees were surveyed about their
observations and preferences. For the final analysis, the amount of
daylight in each store was described as the number of hours per year that
daylight illumination levels exceeded the design electric illumination
level.
Statistical regression models of average
sales for the stores, using up to 50 explanatory variables, and both linear
and natural log descriptions of the variables, found that increased hours of
daylight per store were strongly associated with increased sales, but at a
much smaller magnitude than the previous study. In addition, for this
chain, the daylight effect on sales was found to be constrained by the
amount of parking available at the store site. Sites with parking lots
smaller than the norm experienced decreased sales associated with daylight,
while stores with average and ample parking experienced increased sales as
both the amount of daylight and parking increased. The statistical models
were also more comprehensive, explaining about 75% of the variation in the
data (model R2=0.75), compared to 58% in the previous study.
Specifically, this study found that:
- Average effect of daylighting on sales for
all daylit stores in this chain was variously calculated from 0% to 6%,
depending on the type of model and time period considered.
- A dose/response relationship was found,
whereby more hours of useful daylight per year in a store are associated
with a greater daylight effect on sales.
- No seasonal patterns to this daylight
effect were observed.
- A bound of an empirical daylight effect
for this chain was detailed, with a maximum effect found in the most
favorable stores of about a 40% increase in sales. This upper
bound is consistent with our previous finding.
- Daylight was found to have as much
explanatory power in predicting sales (as indicated by the variable’s
partial R2) as other more traditional measures of retail potential, such
as parking area, number of local competitors, and neighborhood
demographics.
- Along with an increase in average monthly
sales, the daylit stores were also found to have slightly smaller
increase in the number of transactions per month.
- The retailer reported that the primary
motivation for the inclusion of daylight was to save on energy costs by
having photocontrols turn off electric lights when sufficient daylight
was detected. The retailer has been very pleased with the
resulting reduction in operating costs. Based on current energy prices
we estimated average whole building energy savings for the daylit stores
at $0.24/sf for the current design, with a potential for up to $0.66/sf
with a state-of-the art design.
- The value of the energy savings from the
daylighting is far overshadowed by the value of the predicted increase
in sales due to daylighting. By the most conservative estimate, the
profit from increased sales associated with daylight is worth at least
19 times more than the energy savings, and more likely, may be worth
45-100 times more than the energy savings.
- During the California power crisis of
2001, when almost all retailers in the state were operating their stores
at half lighting power, the stores in this chain with daylight were
found to benefit the most, with an average 5.5% increase in sales
relative to the other non-daylit stores within the chain (even while all
stores in this chain increased their sales compared to the previous
period).
- Employees of the daylit stores reported
slightly higher satisfaction with the lighting quality conditions
overall than those in the non-daylit stores. Most strikingly, they
perceived the daylit stores to have more uniform lighting than the
non-daylit stores, even though direct measurements showed both
horizontal and vertical illuminance levels in the daylight stores to be
substantially less uniform.
- Store managers did not report any increase
in maintenance attributable to the skylights.
- The chain studied was found to be saving
about $0.24/sf per year (2003 energy prices) due to use of
photocontrols, which could potentially increase up to $0.66/sf per year
with an optimized daylighting system.
Re-Analysis
Report: Daylighting in Schools, Additional Analysis
– CEC PIER 2001
This report is a follow-on study to the
Daylighting in Schools study[5] that was completed in 1999,
which found a compelling statistical correlation between the amount of
daylighting in elementary school classrooms and the performance of students
on standardized math and reading tests. This re-analysis of the original
study data was intended to answer key questions raised by the peer review of
the earlier study, and expand our understanding of methodological choices
for further work.
The original findings potentially have very
important implications for the design of schools and other buildings where
people live, work and play. Daylight used to be common, and even
required in schools, homes and offices, but fully daylit buildings became
increasingly rare as electric lighting became more the norm. This
re-analysis study helps to provide greater certainty for the original
findings.
For this re-analysis study HMG conducted four
tasks:
- The Teacher Survey collected information
from a sample of teachers in the Capistrano school district about their
education and experience levels, preferences for classroom features and
operation of those features. The primary purpose of the survey was to
provide input to a subsequent "assignment bias" analysis. In
addition, we learned some useful information about teacher preferences,
attitudes and behaviors in response to classrooms conditions.
- While the teachers we surveyed generally
had a preference for windows, daylight and views in their classrooms,
these preferences were not found to be driving classroom
preferences. Far more important was an almost universal desire for
more space, a good location, quiet, lots of storage and water in the
classroom.
- Environmental control was also found to be
an important issue for teachers, especially for those who did not have
full control. Teachers seemed to hold a basic expectation that they
would be able to control light levels, sun penetration, acoustic
conditions, temperature and ventilation in their classrooms. They made
passionate comments about the need for improvement if one or more of
these environmental conditions could not be controlled in their
classroom.
- The Teacher Bias Analysis further examined
information from the Teacher Survey. The survey data was coded into
variables and statistically analyzed in relation to both assignment to
daylit classrooms and the student performance models. The goal of the
Bias Analysis was to discover if the original study had over-inflated
the effect of daylight on student learning by not accounting for a
potential "assignment bias" of better teachers to more daylit
classrooms.
We conclusively found that there was not an
“assignment bias” influencing our results. None of the individual
teacher characteristics we identified were significant in explaining
assignment to a daylit classroom in the Capistrano District. Considering all
teacher characteristics together only explained 1% of the variation in
assignment to daylit classrooms. We did find that a few types of teachers,
those with more experience or honors, were slightly more likely (1%-5%) to
be assigned to classrooms with more windows or some types of skylights.
When we added the teacher characteristics to
the original student performance models, the daylight variables were not
reduced in significance. Further analysis of other sub-populations repeated
these findings. Among twelve models considered, we identified a central
tendency of a 21% improvement in student learning rates from those in
classrooms with the least amount of daylight compared to those with the
most.
In the Grade Level Analysis, we re-analyzed
the original student test score data for both Capistrano and Seattle by
separate grade level, instead of aggregating the data across the four grade
levels (2-5). Our goal was to determine if this method would more
accurately explain the relationship of student performance to daylighting.
We tested for statistical significance and correlation, and we looked at any
patterns discovered in the analysis.
The data did not show any significant
patterns between a daylight effect and the separate grade levels, neither an
increase or decrease in daylight effects by grade level. Thus, we conclude
that there do not seem to be progressive effects as children get older, nor
do younger children seem to be more sensitive to daylight than older
children. Allowing the results to vary by grade did not noticeably improve
the accuracy of the models. Therefore, we conclude that looking at data
across grade levels is a sufficiently accurate methodology.
In the Absenteeism Analysis, we used
absenteeism and tardiness data in the original Capistrano data set as
dependent variables and evaluated them against the full set of explanatory
variables from the original study, plus the new information on teacher
characteristics. These models would allow us to assess whether daylighting
or other classroom physical attributes potentially impacted student health,
as measured by changes in student attendance.
Student attendance data is certainly not the
best indicator of student health. Yet to the extent that attendance data
does reflect student health, our findings do not suggest an obvious
connection between physical classroom characteristics and student health.
Notably, daylighting conditions, operable windows, air conditioning and
portable classrooms were not found to be significant in predicting student
absences.
Overall, the strength of the daylight
variable in predicting student performance stands out sharply across all of
these re-analysis efforts.
This analysis also demonstrated that the
findings of these models are more strongly dependent upon the sample
population then the subtleties of the explanatory variables. Thus, we
believe that it will be more informative to replicate this study with a
different population, to continue to try to refine the models with further
detail in the explanatory variables.
Daylighting
in Schools – PG&E 1999
An Investigation into the
Relationship between Daylighting and Human Performance
This study looks at the effect of daylighting
on human performance. It includes a focus on skylighting as a way to isolate
daylight as an illumination source, and separate illumination effects from
other qualities associated with daylighting from windows. In this project,
we established a statistically compelling connection between daylighting and
student performance, and between skylighting and retail sales. This report
focuses on the school analysis.
We obtained student performance data from
three elementary school districts and looked for a correlation to the amount
of daylight provided by each student’s classroom environment. We used data
from second through fifth grade students in elementary schools because there
is extensive data available from highly standardized tests administered to
these students, and because elementary school students are generally
assigned to one teacher in one classroom for the school year. Thus, we
reasoned that if the physical environment does indeed have an effect on
student performance, we would be mostly likely to be able to establish such
a correlation by looking at the performance of elementary school students.
We analyzed test score results for over
21,000 student records from the three districts, located in Orange Country,
California, Seattle , Washington, and Fort Collins, Colorado. The data sets
included information about student demographic characteristics and
participation in special school programs. We reviewed architectural plans,
aerial photographs and maintenance records and visited a sample of the
schools in each district to classify the daylighting conditions in over 2000
classrooms. Each classroom was assigned a series of codes on a simple 0-5
scale indicating the size and tint of its windows, the presence and type of
any skylighting, and the overall amount of daylight expected.
The study used multivariate linear regression
analysis to control for other influences on student performance. Regressions
were compared using data from two separate tests, math and reading, for each
district. Each math and reading model was also run separately using first
the window and skylight codes, and then the overall daylight code. We
reasoned that if daylight effects were truly robust the variables should
perform similarly in all models. Thus, we created a total of twelve models
for comparison, consisting of four models for each of three districts.
The daylighting conditions at the Capistrano
school district were the most diverse, and the data from that district were
also the most detailed. Thus Capistrano became our most precise model. In
this district, we were able to study the change in student test scores over
a school year. Controlling for all other influences, we found that students
with the most daylighting in their classrooms progressed 20% faster on math
tests and 26% on reading tests in one year than those with the least.
Similarly, students with the largest window areas were found to progress15%
faster in math and 23% faster in reading than those with the least. And
students that had a well-designed skylight in their room, one that diffused
the daylight throughout the room and which allowed teachers to control the
amount of daylight entering the room, also improved by 19-20% faster than
those students without a skylight. We also identified another window-related
effect, in that students in classrooms where windows could be opened were
found to progress 7-8% faster than those with fixed windows, regardless of
whether they also had air conditioning. These effects were all observed with
99% statistical certainty.
The studies in Seattle and Fort Collins used
the final scores on math and reading tests at the end of the school year,
rather than the amount of change from the beginning of the year. In both of
these districts we also found positive, and highly significant, effects for
daylighting. Students in classrooms with the most daylighting were found to
have 7% to 18% higher scores than those with the least.
The three districts have different curriculum
and teaching styles, different school building designs and very different
climates. And yet the results of studies show consistently positive and
highly significant effects. This consistency persuasively argues that there
is a valid and predictable effect of daylighting on student performance.
The results of this study of student
performance, when combined with the companion study showing the positive
effect of skylighting on retail sales, also strongly support the thesis that
these performance benefits from daylighting can be translated to other
building types and human activities.
Skylighting
and Retail Sales – PG&E 1999
An
Investigation into the Relationship between Daylighting and Human
Performance
This study looks at the effect of daylighting
on human performance. It specifically focuses on skylighting as a way to
isolate daylight as an illumination source, and avoid all of the other
qualities associated with daylighting from windows. In this project, we
established a statistically compelling connection between skylighting and
retail sales, and between daylighting and student performance. This report
focuses on the retail analysis.
We analyzed data on the sales performance of
a chain retailer who operates a set of nearly identical stores. The analysis
included 108 stores, where two thirds of the stores have skylighting and one
third do not. The design and operation of all the store sites is remarkably
uniform, with the exception of the presence of skylights in some. The
electric lighting was primarily fluorescent. The skylights often provided
far more illumination, often two to three times the target illumination
levels. Photo-sensor controls turned off some of the fluorescent lights when
daylight levels exceeded target illumination.
The monthly gross sales per store were
averaged over an 18-month period that went from February 1 of one year to
August 31 of the following year. This average sales figure was transformed
into a "sales index" that we could manipulate statistically, but
that did not reveal actual dollar performance. Stores in the sample were
selected to operate within a limited geographic region that had similar
climatic conditions, and to have a constrained range of size and age. The
geographic region has a relatively sunny climate. All of the stores in the
data set are one story.
The multivariate regression analysis allowed
us to control for the influence of other variables, which might influence
sales. Other variables considered included the size and age of the store,
hours of operation, and economic characteristics associated with the zip
code location.
Skylights were found to be positively and
significantly correlated to higher sales. All other things being equal, an
average non-skylit store in the chain would be likely to have 40% higher
sales with the addition of skylights, with a probable range somewhere
between 31% to 49%. This was found with 99% statistical certainty. After the
number of hours open per week, the presence of skylights was the best
predictor of the sales per store of all the variables that we considered.
Thus, if a typical non-skylit store were averaging sales of $2/sf, then its
sales might be expected to increase to somewhere between $2.61 to $2.98 with
the addition of a skylighting system.
The skylights are seen to have a major impact
on the overall operation of the chain. Were the chain to add the skylighting
system to the remaining 33% of their stores, their yearly gross sales are
predicted to increase by 11%. The difference between having none of their
stores skylit and all their stores skylit is a 40% increase in gross sales
for the retail chain.
Footnotes:
|
