# Classroom acoustics affect student achievement

## Classroom mechanical systems should be designed with lower noise levels to optimize student reading comprehension.

Low noise levels and controlled reverberation are commonly accepted as requirements for good speech intelligibility in classrooms. If the background noise in classrooms is too loud, the signal-to-noise ratio may not be high enough for students to separate speech from noise. This may lead to decreased comprehension of speech or reduced speech intelligibility. High reverberation times can also smear speech signals, causing them to be unintelligible.

The current ANSI Standard S12.60-2010 specifies a maximum unoccupied background noise level (BNL) of 35 dBA and 55 dBC for core learning spaces with single-mode HVAC systems. The standard calls for a maximum BNL of 37 dBA and 57 dBC for classrooms with multiple-mode HVAC systems.

An A-weighted BNL (dBA) reduces sound energy in the lower frequencies to better relate to how humans perceive sound, since humans are more sensitive to mid and high frequencies. A C-weighted BNL (dBC) applies a fairly flat weighting to all frequencies. The maximum unoccupied reverberation time (RT) allowed by the ANSI Standard S12.60-2010 is 0.6 sec in each of the 500, 1,000, and 2,000 Hz octave bands for core learning spaces smaller than 9,994 cu ft. The requirements in this standard are based on achieving signal-to-noise ratios necessary for good speech intelligibility.

However, these ANSI requirements have not been well-linked to student achievement. Previous research has shown that noise generated in occupied classroom environments can be significantly related to achievement for young children. Building standards typically specify unoccupied classroom acoustical conditions, since these may be dictated through the building design phase. This study investigates how unoccupied BNLs and RTs relate to elementary student achievement.

**Study methods**

This research was conducted in 58 elementary school classrooms in the Council Bluffs Community School District in Council Bluffs, Iowa. This encompassed all of the second- and fourth-grade classrooms in the district during the 2008/2009 academic year. These classrooms all had closed floor plan designs, wherein all doors and windows to adjacent spaces could be closed. Typical room finishes included acoustical ceiling tile, thin floor carpet, and hard wall surfaces. Figure 1 shows views of a typical classroom tested with similar room layouts and furnishings. The classroom ceilings were 8 to 10 ft high with floor areas of approximately 750 sq ft. All of the classrooms surveyed had volumes less than 9,994 cu ft. The original construction dates and the most recent renovation or addition dates for the elementary schools are shown in Table 1. Acoustical measurements, including BNL and RT, were gathered in the unoccupied classrooms between April and June 2009. The unoccupied BNLs and RTs were then compared to the standardized achievement scores from students in the surveyed classrooms.

**Acoustical measurement procedures**

The unoccupied BNLs were recorded as A-weighted (L_{Aeq}) and C-weighted (L_{Ceq}) equivalent noise levels using a Type 1 sound level meter over a 5-min time period. Most of the classrooms had central mechanical systems that were operating during the BNL measurements. However, two of the classrooms had central mechanical systems that were deactivated for the BNL measurements. Therefore, the classrooms in this school (School L) are omitted from the BNL results shown. Also, six of the classrooms had window air-conditioning units that were activated during the BNL measurements (Classrooms M1, M2, M3, N2, N3, and N6). The RT was measured in each unoccupied classroom from an impulse response generated by popping a balloon and recorded by a Type 1 sound level meter.

**Student achievement tests**

The Iowa Test of Basic Skills, including reading comprehension and math subject areas, was administered to the students in the surveyed classrooms during April 2009. The results showed the percentage of proficient students averaged per grade level per school. For the 2008/2009 academic year, any score above the 41^{st} percentile was defined by the state of Iowa as a proficient score. Also, poverty rates were collected to control for demographic differences occurring among the students. These were reported as the percentage of students who received a fee waiver or reduction of fees, averaged per school. This information was provided by the school district. Students who lived in households below a certain income level were eligible to apply for a fee waiver or reduction. For a family of four, the gross income could not exceed $28,665 to receive a fee waiver and the gross income could not exceed $40,793 to receive a fee reduction. These income levels were based on U.S. Dept. of Agriculture (USDA) requirements.

**Results**

**Background noise level**

The A-weighted (L_{Aeq}) and C-weighted (L_{Ceq}) equivalent noise levels were measured in the unoccupied second-grade and fourth-grade classrooms (Figures 2 and 3). Classrooms with the same letter designation were located in the same school. All of the classrooms have BNLs higher than the 35 dBA specified in the ANSI classroom acoustics standard, and most of the BNLs are greater than the allowable L_{Ceq} value of 55 dBC. However, the difference between the A-weighted and C-weighted equivalent noise levels is less than or equal to 22 dB for all of the classrooms, indicating that the low-frequency noise was not excessively dominant.

Because the student achievement scores were only available as grade-level averages per school, the L_{Aeq} values were also averaged per grade level per school. The L_{Aeq} values for the 11 schools with consistent HVAC conditions (activated central mechanical systems) during the BNL measurements are shown in Figure 4. The bars on the graph in Figure 4 show the range in L_{Aeq} values about the average value for all classrooms at each grade level in each school.

The study also assessed the impact of the age of the school buildings on the classroom noise levels. The original school construction dates are significantly correlated to both the L_{Aeq} values (*r* = -0.46, *p* < 0.05) and L_{Ceq} values (*r* = -0.48, *p* < 0.05), meaning that older schools tended to have higher noise levels (Figure 5). Schools constructed more recently typically had lower unoccupied BNLs. The most recent school renovation or addition dates are not significantly correlated to the unoccupied classroom BNLs; this may be due to the fact that those renovations did not necessarily impact the mechanical system serving the classrooms.

**Reverberation time**

The unoccupied classroom RTs calculated from the balloon pop impulse response measurements are shown in Figure 6. This figure shows the RT value in each classroom averaged across the 500 and 1,000 Hz octave bands. The mid-frequency RTs range from 0.2 to 0.6 sec, which are all less than the maximum RT of 0.6 sec recommended in the ANSI S12.60 Standard.

**Student Achievement Scores and Poverty Rates**

The results from the standardized student achievement scores and poverty rates from the 2008/2009 academic year are shown in Table 1. The state of Iowa set the minimum desired performance trajectory for the fourth-grade students to be 76% proficient for reading comprehension and 74.7% proficient for math.

**Analyzing the data**

**Statistical analysis methods**

Correlation and regression analyses were performed on the data to determine what relationships exist, if any, between the acoustics variables and the standardized student achievement scores. A correlation indicates the strength and nature (positive, negative, or non-existent) of the linear relationship between two variables. One method to determine the correlation between two variables is to calculate the Pearson product-moment correlation coefficient, *r*. This coefficient ranges from -1 to +1. Negative coefficients indicate that there is an inverse or negative correlation between the variables, whereas positive coefficients occur for direct or positive correlations between two variables.

The significance of the correlation is determined by the probability that the correlation would have occurred due to chance alone, or the *p*-value. Typically, if the *p*-value is less than 0.05, the correlation is considered statistically significant. For example, if the *p*-value is less than 0.05, then the probability that the correlation coefficient occurred for the sample due to chance alone is less than 0.05. This indicates that the relationship found in the sample data set is likely to occur in the general population.

Regressions quantify the amount of variance in one variable that is due to the variance in one or more other variables. A simple linear regression determines the linear relationship between a dependent variable and one predictor variable. To determine how much variance in the dependent variable is accounted for by the predictor variable, an *R ^{2}* value is calculated. If the

*R*value is multiplied by 100, it is the percentage of variance accounted for in the dependent variable by the predictor variable. If the

^{2}*p*-value associated with

*R*is less than 0.01, then there is a very small chance that the percentage of variance accounted for by the regression model is due to chance alone.

^{2}**Background noise level plays significant role**

The unoccupied average L_{Aeq} values per grade level for the 11 schools with central mechanical systems that were activated during the BNL measurements, shown in Figure 4, were compared to the standardized student achievement scores. The correlation between BNL and math is nonsignificant. However a significant negative correlation occurs between L_{Aeq} and reading comprehension (*r* = -0.55, *p* < 0.01). When controlling for the effects of poverty rates on the reading comprehension scores, the semipartial correlation value of -0.49 between L_{Aeq} and reading comprehension is also significant (*p* < 0.05). These results indicate that higher reading comprehension scores occur in classrooms with lower unoccupied A-weighted noise levels.

Regression analyses with L_{Aeq} as a predictor variable for the reading comprehension scores were also conducted. A scatter plot between the L_{Aeq} values and the reading comprehension scores along with regression equations calculated for the classrooms are shown in Figure 7. The regression equation calculated for the combined second- and fourth-grade classrooms is significant, with BNL accounting for 30% of the variance in the reading comprehension scores (*R ^{2}* = 0.30,

*p*< 0.01). The regression equation for the fourth-grade classrooms indicates that the highest acceptable classroom BNL is 41 dBA to meet the state trajectory of having 76% of the fourth-graders proficient in reading comprehension.

**Reverberation time has no effect**

Correlations relating the mid-frequency RTs shown in Figure 6 to the standardized student achievement scores were calculated. The unoccupied RTs are not significantly correlated to either the reading comprehension or math student achievement scores. Therefore, results from further statistical analyses including RT are not shown. This result may not be surprising, though, since all of the classrooms surveyed did meet the RT requirements set forth by ANSI S12.60.

**Conclusion**

This study compared unoccupied classroom noise levels and reverberation times to the age of the school buildings and the elementary student achievement scores attained by students using those classrooms. All of the classrooms surveyed had volumes less than 9,994 cu ft, for which the ANSI S12.60 Standard recommends a maximum RT of 0.6 sec. The unoccupied RTs were not significantly correlated to student achievement, possibly due to the fact that no RT values higher than 0.6 sec were measured. Continuing research should include classrooms with longer RTs to determine if the RT limit specified in the standard is too low.

Both the A-weighted and C-weighted noise equivalent noise levels in classrooms with the central HVAC systems operating were significantly negatively correlated to the original school building construction dates. This suggests that classrooms constructed more recently tend to have lower unoccupied BNLs.

Also, a significant negative correlation occurred between the A-weighted equivalent noise levels and student reading comprehension, even when controlling for the effects of poverty rates on achievement. The results indicate that classrooms should be designed with unoccupied BNLs less than 41 dBA to meet the minimum desired Iowa state reading comprehension performance trajectory, though the reading comprehension scores generally continued to improve for classrooms with BNLs less than 41 dBA. The current ANSI S12.60 Standard recommends an upper limit of 35 dBA for classrooms with single mode HVAC systems.

This study does not suggest the BNL limits specified in the ANSI standard should be increased. Rather, continuing research in this area is recommended; suggestions are to include measurements in classrooms with RTs above 0.6 sec and with the mechanical systems operating in various modes.

*Ronsse received her PhD from the architectural engineering program at the University of Nebraska-Lincoln. She is an active member of the Acoustical Society of America (ASA) and ASHRAE, currently serving as the Standards Subcommittee Chair for the ASHRAE Technical Committee 2.6 on Sound and Vibration Control. Wang is an associate professor of architectural engineering in the Durham School of Architectural Engineering and Construction at the University of Nebraska-Lincoln. She is a fellow of the ASA, board-certified by the Institute of Noise Control Engineering, and the current chair of ASHRAE's Technical Committee 2.6 on Sound and Vibration Control.*

**References**

- ANSI/ASA S12.60-2010/Part 1. 2010. Acoustical Performance Criteria, Design Requirements, and Guidelines for Schools, Part 1: Permanent Schools. New York: American National Standards Institute.
- Bistafa, S.R. and Bradley, J.S. 2000. Reverberation time and maximum background noise level for classrooms from a comparative study of speech intelligibility metrics. J. Acoust. Soc. Am., 107 (2): 861 - 875.
- Bradley, J.S. 1986. Speech intelligibility studies in classrooms. J. Acoust. Soc. Am., 80 (3): 846 - 854.
- Bradley, J.S. and Sato, H. 2008. The intelligibility of speech in elementary school classrooms. J. Acoust. Soc. Am., 123 (4): 2078 - 2086.
- Elliott, L.L. 1979. Performance of children aged 9 to 17 years on a test of speech intelligibility in noise using sentence material with controlled word predictability. J. Acoust. Soc. Am., 66 (3): 651 – 653.
- Field, A. 2000.
*Discovering Statistics Using SPSS for Windows*. London: Sage Publications, Ltd. - Field, A. and Hole, G. 2003.
*How to Design and Report Experiments*. London: Sage Publications, Ltd. - Shield, B.M. and Dockrell, J.E. 2008. The effects of environmental and classroom noise on the academic attainments of primary school children. J. Acoust. Soc. Am., 123 (1): 133 – 144.

- Events & Awards
- Magazine Archives
- Oil & Gas Engineering
- Supplements
- Salary Survey
- Blogs
- Digital Reports
- Survey Prize Winners

### Annual Salary Survey

Before the calendar turned, 2016 already had the makings of a pivotal year for manufacturing, and for the world.

There were the big events for the year, including the United States as Partner Country at Hannover Messe in April and the 2016 International Manufacturing Technology Show in Chicago in September. There's also the matter of the U.S. presidential elections in November, which promise to shape policy in manufacturing for years to come.

But the year started with global economic turmoil, as a slowdown in Chinese manufacturing triggered a worldwide stock hiccup that sent values plummeting. The continued plunge in world oil prices has resulted in a slowdown in exploration and, by extension, the manufacture of exploration equipment.

**Read more: 2015 Salary Survey**