Grantee Research Project Results
Final Report: Evidence-Based Interactions between Indoor Environmental Factors andTheir Effects on K-12 Student Achievement
EPA Grant Number: R835633Title: Evidence-Based Interactions between Indoor Environmental Factors andTheir Effects on K-12 Student Achievement
Investigators: Wang, Lily , Bovaird, James , Lau, Josephine , Waters, Clarence
Institution: University of Nebraska at Lincoln
EPA Project Officer: Hahn, Intaek
Project Period: November 1, 2014 through October 31, 2018 (Extended to October 31, 2019)
Project Amount: $998,433
RFA: Healthy Schools: Environmental Factors, Children’s Health and Performance, and Sustainable Building Practices (2013) RFA Text | Recipients Lists
Research Category: Children's Health , Human Health
Objective:
The proposed research aims to establish how indoor environmental conditions in K-12 school buildings impact student scholastic achievement. The objectives are: (1) to study comprehensively the impacts of a wide set of indoor environmental factors (including indoor air quality, thermal, lighting, and acoustic conditions) on student achievement; (2) to investigate how these conditions interact with each other to impact student achievement; (3) to rank order the environmental variables in terms of their relative impact on student achievement; and (4) to determine how these effects vary with different demographic (e.g. socio-economic) groups.
Summary/Accomplishments (Outputs/Outcomes):
Information about the indoor air quality (IAQ), thermal comfort, lighting, and acoustic conditions have been collected from 220 classrooms across five school districts in Nebraska and Iowa. Data were collected under occupied and unoccupied conditions for two days in three seasons from 2015-2017. IAQ and thermal measurements included the indoor concentration of carbon dioxide, formaldehyde, the count of particles with aerosol diameters ranging from 0.3 µm to 2.5 µm and aerosol diameters ranging from 2.5 µm to 10 µm, air temperature, relative humidity, and globe temperature. View, daylighting and electric lighting data were collected to understand lighting conditions. Daylighting models were created and annual hourly simulations were conducted with actual meteorological year (AMY) weather data to create daylighting metrics. Assorted background noise levels and room impulse responses from which reverberation times are extrapolated were collected for acoustics data. In addition, (1) classroom-averaged student achievement data on annual standardized reading and math tests and (2) classroom-aggregated demographics such as the percent of students in each class receiving free or reduced lunch, the percent of gifted students in each class, and the percent of special education students in each class were compiled for this study.
The field measurements revealed that all classrooms meet IES recommended illuminance level for reading and writing but only 20% of classrooms in this study meet the ASHRAE Std. 62.1 ventilation rate requirements. In comparison to ANSI S12.60, 91% of the classrooms do not meet the recommended maximum background noise level for unoccupied conditions, while 15% do not meet the recommended maximum reverberation time.
The team conducted comprehensive statistical analyses on the data collected from these 220 classrooms to examine the relationships between environmental conditions and student achievement. Structural equation models were constructed and tested for each technical discipline (acoustics, lighting, and thermal and indoor air quality). Issues were identified with the thermal and indoor air quality models that required more in-depth exploration as to which metrics had to be considered seasonally, rather than as one aggregate metric for the year.
In the end, our team concluded that it was not possible to construct structural equation models for the lighting data, nor for the thermal and indoor air quality data. So instead, we used multivariate regression models to test for significant relationships. Because there were a large number of predictors (or measured environmental metrics), we first studied each technical discipline on its own (acoustics model alone, lighting model alone, thermal and indoor air quality model alone). Then all three models were combined into a cumulative model.
ACOUSTICS MODEL
The final acoustics structural equation model had three main predictive latent variable constructs: one related to levels of speech in the classroom, one related to levels of non-speech in the classroom, and classroom reverberation time. When analyzed against student achievement score while controlling for demographics, the model showed that math achievement scores were statistically significantly related to the levels of speech levels in classrooms; louder speech levels corresponded to lower math achievement scores, after controlling for other classroom-aggregated student demographics. Later, a multivariate regression model that used the metric LAeq,speech(or the A-weighted equivalent sound level when speech was occurring in the classroom over a class day) in place of the speech level latent construct, and L90non-speech(or the A-weighted sound level exceeded 90% of the time when no speech was occurring in the classroom over a class day) in place of the non-speech level latent construct, demonstrated similar results.
LIGHTING MODEL
The multivariate regression model for the lighting conditions originally contained several metrics that analyzed daylight conditions, electric illuminance conditions and view conditions of classrooms. Daylight metrics were made through computer simulation of each classroom. This created a problem of non-independence as only one daylight metric could be chosen. This same problem occurred for the electric illuminance metrics. In the end, the multiple regression model that was created contained a daylight glare potential metric, a time-weighted electric illuminance level, a horizontal view angle (quantifies the amount of windows in a space), and a view layers metric (quantifies what can be seen outside of the windows). The multivariate linear regression found that more view layers corresponded to higher reading achievement scores, after controlling for other classroom-aggregated student demographics. When students had views of the ground, the landscape, and the sky, they performed better than students with no view or view to just one or two of the layers.
THERMAL AND INDOOR AIR QUALITY MODEL
Since there were observable seasonal variations in the gathered thermal/IAQ data, an omnibus test was conducted to decide if the gathered thermal/IAQ metrics varied significantly across seasons, or if seasonal variations were not significant so that values could be averaged into one annual measure. Based on the analyses, the following data were considered with seasonal variation: air temperature, globe temperature, fine particles, formaldehyde, TVOC, velocity, and ventilation rate. Meanwhile, data for carbon monoxide, nitrogen dioxide, ozone, coarse particles, relative humidity were averaged into annual values.
The multivariate regression model for the thermal and indoor air quality conditions showed that math achievement scores were statistically significantly related to the globe temperature in classrooms while controlling for demographics. During the winter seasons, higher globe temperature corresponded to higher math achievement scores, while for spring seasons, lower globe temperature corresponded to higher math achievement scores.Additionally, higher air temperatures in the winter were associated with increased math scores.These trends may be explained as occupants having different thermal comfort preferences across different seasons.
Other findings are that increased formaldehyde exposure (in the fall season) was associated with the reduced math scores. In comparison to classrooms with centralized air systems, classrooms with unit ventilators were associated with reduced math scores.
For reading achievement, higher air temperatures in the winter were associated with increased reading scores, as also seen with math scores. Increased air velocity in the winter was associated with reduced reading scores,while increased ventilation rate in the fall was associated with higher reading achievement. Ozone levels, PM_2.5 concentration in winter and TVOC in winter are other factors that were found to be statistically significantly related to reading scores.
COMBINED MODEL
A combined model with acoustic, lighting, thermal, and indoor air quality metrics was analyzed to understand how these conditions may interact and whether or not the same results from the individual models would still be visible. This resulted in a large model with 34 independent variables. Results indicate that no acoustics or lighting metrics were found to have significant effects in the overall model, while the thermal comfort and indoor air quality metrics showed similar results on student achievement as the individual model.
INDIVIDUAL TECHNICAL DISCIPLINE MODELS WITH DEMOGRAPHIC INTERACTIONS
To determine how the relationship between environmental conditions and student achievement may vary with different demographic groups, we started by testing each individual technical discipline model for statistically significant interactions with the collected demographics. The following results were found.
Acoustics
Two statistically significant interactions were found:
As the percent of free and reduced lunch recipients in a classroom increases, a higher level of speech in a classroom has a greater negative relationship with math achievement scores. Holding all other variables constant at their mean value, Fig. 1 shows an example of this relationship at the mean percent of free and reduced lunch recipients, and then at one standard deviation away from the mean.
Figure 1. Interaction of LAeq(s) with Percent Free and Reduced Lunch Recipients on Math Achievement
As the percent of gifted students in a classroom increases, higher levels of non-speech in a classroom has greater positive relationship with reading achievement scores.
Lighting
One statistically significant interaction was found:
As the percent of free and reduced lunch recipients in a classroom increases, a higher horizontal sight angle in a classroom has a greater negative relationship with reading achievement scores. Holding all other variables constant at their mean value, Fig. 2 shows an example of this relationship at the mean percent of free and reduced lunch recipients, and then at one standard deviation away from the mean.
Figure 2. Interaction of Horizontal Sigh Angle with Percent Free and Reduced Lunch Recipients on Reading Achievement
Thermal and Indoor Air Quality
Ten statistically significant interactions were found. Of these, the following are a few that we are highlighting in this report as being of most interest to the community:
As the percent of free and reduced lunch recipients in a classroom increases, higher TVOC (in winter season) has a greater negative relationship with reading achievement scores. Holding all other variables constant at their mean value, Fig. 3 shows an example of this relationship at the mean percent of free and reduced lunch recipients, and then at one standard deviation away from the mean.
As the percent of gifted students in a classroom increases, higher ventilation rate (in spring season) has a greater positive relationship with reading achievement scores. Holding all other variables constant at their mean value, Fig. 4 shows an example of this relationship at the mean percent of gifted students, and then at one standard deviation away from the mean.
Figure 4b. Interaction of Ventilation(s) with Gifted Students on Reading Achievement
As the percent of special education students in a classroom increases, higher globe temperature (in winter season) has a greater positive relationship with math achievement scores. Holding all other variables constant at their mean value, Fig. 5 shows an example of this relationship at the mean percent of students receiving special education, and then at one standard deviation away from the mean.
Figure 5. Interaction of Globe Temperature (W) and Special Education on Math Achievement
As the percent of special education students in a classroom increases, ventilation rate (in spring season) has a greater negative relationship with math achievement scores. Holding all other variables constant at their mean value, Fig. 6 shows an example of this relationship at the mean percent of students receiving special education, and then at one standard deviation away from the mean.
Figure 6. Interaction on Ventilation (S) on Special Education Math Achievement
The team is still working on combining the three technical area models into one combined model with demographic interactions. The combined model ends up having a large number of variables that must be dealt with carefully.
In Year 4 of this project, additional measurements were made in 55 classrooms towards a goal of exploring potential causal relationships between environmental conditions and student achievement. However, many of our partnering school districts were unable to make the degree of changes we recommended to the environmental systems in their classrooms. Only 21 of the 55 classrooms may be analyzed to directly compare results before and after a change was made. Based on what we learned from analyzing the data gathered in Years 2 and 3, the data gathering process in Year 4 lasted four days instead of two, and we only measured once in heating season and once in cooling season, rather than in the three seasons (fall, winter, spring). Over the past year, the Year 4 data have been processed but we have not had a chance to run the causal analyses or cross-validation study of results from Years 2 and 3.
We have continued to engage with the design and school communities, to share the results of our work. We have met with leaders from each school district partner in this study and shared data specific to their schools. We are preparing a condensed report of our findings to share with those districts that highlights how indoor environmental quality may be optimized to benefit occupants in educational settings, so that our school partners may also report out to their constituencies. Furthermore, we have continued to present on our findings at professional conferences related to the K-12 building engineering and facility management fields.
Conclusions:
While the EPA funding for this project has officially come to an end, we have many more plans to study and learn from the gathered data sets. We continue to finalize the combined model with demographic interactions, and to finish analyses of the Year 4 data in which causal relationships will be studied for the 21 classrooms in which a change was made, while data from the other 34 additional classrooms re used to cross-validate the model developed from the 220-classroom sample. Finally, we are finalizing drafts of at least four more manuscripts for peer-reviewed publication, based on the data and analyses from this project.
Journal Articles on this Report : 5 Displayed | Download in RIS Format
Other project views: | All 24 publications | 9 publications in selected types | All 9 journal articles |
---|
Type | Citation | ||
---|---|---|---|
|
Deng S, Zou B, Lau J. The Adverse Associations of Classrooms’ Indoor Air Quality and Thermal Comfort Conditions on Students’ Illness Related Absenteeism between Heating and Non-Heating Seasons—A Pilot Study. International Journal of Environmental Research and Public Health 2021;18(4):1500. |
R835633 (Final) |
Exit Exit |
|
Kabirikopaei A, Lau J. Uncertainty analysis of various CO 2-Based tracer-gas methods for estimating seasonal ventilation rates in classrooms with different mechanical systems. Building and Environment 2020;179(107003). |
R835633 (Final) |
Exit Exit |
|
Kabirikopaei A, Lau J, Nord J, Boviard J. Identifying the K-12 classrooms' indoor air quality factors that affect student academic performance. Science of The Total Environment 2021;786. |
R835633 (Final) |
Exit Exit |
|
Kuhlenengel M, Konstantzos I, Waters C. The Effects of the Visual Environment on K-12 Student Achievement. Buildings 2021;11(11). |
R835633 (Final) |
Exit Exit |
|
Wang L, Brill L. Speech and noise levels measured in occupied K–12 classrooms. The Journal of the Acoustical Society of America 2021;150(2). |
R835633 (Final) |
Exit Exit Exit |
Supplemental Keywords:
student achievement, indoor environmental quality, indoor air quality, thermal conditions, lighting conditions, acoustic conditions, green buildings, high performance buildings, cost-benefit, architectural engineering, mixed linear models, midwest, Nebraska (NE), Iowa (IA)Relevant Websites:
School Environmental Effects on Student Achievement (SEESA) Exit
Progress and Final Reports:
Original AbstractThe perspectives, information and conclusions conveyed in research project abstracts, progress reports, final reports, journal abstracts and journal publications convey the viewpoints of the principal investigator and may not represent the views and policies of ORD and EPA. Conclusions drawn by the principal investigators have not been reviewed by the Agency.
Project Research Results
- 2018 Progress Report
- 2017 Progress Report
- 2016 Progress Report
- 2015 Progress Report
- Original Abstract
9 journal articles for this project