Abstract
Objective: Previous research has concluded that teachers are at a higher-than-normal risk for voice issues that can cause occupational limitations. While some risk factors have been identified, there are still many unknowns. Patients and Methods: A survey was distributed electronically with 506 female teacher respondents. The survey included questions to quantify three aspects of vocal fatigue as captured by the Vocal Fatigue Index (VFI): (1) general tiredness of voice (performance), (2) physical discomfort associated with voicing (pain), and (3) improvement of symptoms with rest (recovery). The effect of classroom capacity on US teachers’ self-reported experience of vocal fatigue was analyzed. Results: The results indicated that a classroom’s capacity significantly affected teachers’ reported amounts of vocal fatigue, while a teacher’s age also appeared to significantly affect the reported amount of vocal fatigue. A quadratic rather than linear effect was seen, with the largest age effect occurring at around 40–45 years in all three factors of the VFI. Conclusion: Further factors which may affect vocal fatigue must be explored in future research. By understanding what increases the risk for vocal fatigue, educators and school administrators can take precautions to mitigate the occupational risk of short- and long-term vocal health issues in school teachers.
Introduction
Over the last three decades, there has been a heightened interest in examining voice problems associated with people who use their voice as a tool of the trade, such as school teachers. Research has indicated that school teachers are at high risk for experiencing short- and long-term problems with their voice [1, 2]. For example, in a survey of over 2,500 teachers and non-teachers in the USA, it was reported that 57.7% of school teachers had experienced voice problems during their lifetimes. In comparison, only 28.8% of non-teachers had experienced voice problems [3] during their lives. According to the study, 11% of the school teacher population surveyed also reported having current voice issues (almost double that of non-teachers). Additionally, female occupational voice users have been reported to have a higher instance of voice problems than males [4]. In these cases, not only is the teacher’s health and quality of life impacted, but the students are less likely to learn the information taught in class [5, 6].
A common vocal complaint of teachers is vocal fatigue. Vocal fatigue has been described clinically as an adverse reaction of voice users to extended periods of vocal load [7]. Titze et al. [8] has suggested that symptoms of vocal fatigue may be reduced by adding small short-term periods of vocal silence throughout the day. Additionally, other research has indicated that hours or even days may be needed in order to fully recover from multiple hours of prolonged voice use [9]. Nevertheless, vocal fatigue has been difficult to quantify, which has made it difficult to directly link potential risk factors to voice problems.
Recently, a Vocal Fatigue Index (VFI) [10] was validated which may aid in providing a more standard quantifying of vocal fatigue. The VFI is a 19-question scale used to quantify the amount of vocal fatigue experienced by a given individual in the following three scored categories: (1) general tiredness of voice (performance), (2) physical discomfort associated with voicing (pain), and (3) improvement of symptoms with rest (recovery). The VFI may be a useful tool for identifying those experiencing vocal fatigue as well as factors that affect the amount vocal fatigue reported by occupational voice users.
Many different factors have been implicated as possible contributors to the increased vocal fatigue and vocal risk in occupational voice users. One such factor is classroom acoustics. Poor classroom acoustics have been shown to be a factor that may increase the likelihood that school teachers experience voice problems. Bottalico et al. [11] indicated that in classrooms with longer reverberation times (generally associated with larger volumes) teachers adopt vocal behaviors that are increasing the likelihood of acquiring voice problems. Kob et al. [12] reported that differences in room acoustical conditions in classrooms seemed to affect teachers with voice problems more than healthy teachers. Therefore, vocal health issues could likely be improved by improving classroom acoustics. This could be associated with the fact that more time speaking in larger and more reverberant classrooms increases reports of vocal effort.
Brunskog et al. [13] objectively investigated vocal adjustments and subjective judgments of talkers due to changes in the physical attributes of a room (e.g., size and reverberation time). An increase in a talker’s vocal power was found to be correlated with room size and the amplification by the room of the talker’s voice at her/his ears (termed “room gain”).
High noise levels in classrooms are a frequently mentioned cause of teachers’ voice problems, with noise levels varying between 56.3 and 76.8 dB(A) according to the children’s activities and the number of children in the classroom [14, 15]. As classroom dimensions increase to accommodate more children, the noise present in those active classrooms increases as well. Though regulations differ from state to state in the USA, there are recommendations for the number of students allowed per classroom and per teacher. These recommendations are calculated by rule and are roughly 20 square feet of the total square footage of a classroom for each student (National Association of State Fire Marshals).
In 2002, the American National Standards Institute published guidelines regarding optimal classroom acoustics (ANSI S12.60-2002). The ANSI recommended that the background noise present in an unoccupied classroom remain at or below 35 dB(A). They also recommended that reverberation times remain <0.6 s in smaller classrooms and <0.7 s in larger classrooms. Knecht et al. [16] found that most school classrooms surveyed (87.5%) did not meet either of these requirements.
Additionally, some research has suggested that a teacher’s age may have a significant effect on her/his reported amounts of voice problems including vocal fatigue. More specifically, teachers older than 40 years have reported an increased risk of experiencing voice disorders compared to those younger than 40 years [3]. This increased risk may be associated with hormonal and other physiological changes that occur with age. For example, hyaluronic acid in the vocal folds, which functions as a shock absorber, viscosity regulator, and important signaling molecule, is different between genders and decreases with age [17, 18]. These and other age-related changes may affect the stiffness of the vocal folds and thus potentially increase the number of voice problems experienced by teachers.
Given the elevated vocal risk of teachers generally, and the likely tie between vocal risk and poor classroom acoustics, the current study was designed to answer the following question: how is reported vocal fatigue (as quantified by the VFI) in an average female teacher associated with classroom size when controlling for the age of the teacher? In addressing this question, the hypothesis was that female teachers would report higher levels of vocal fatigue when (1) teaching in larger classrooms and (2) their age was increased. While age is not the primary issue, it should be accounted for, since the voice does change with age [17] and a teacher’s age has been found to have a negative effect on the reported amount of voice problems, especially in females. For example, several studies have found that females older than 40 years consistently experience more instances of voice problems than their male counterparts [3, 18, 19].
Subjects and Methods
Subjects
This research was conducted in accordance with protocols outlined by Michigan State University’s Human Research Protection Program (determined exempt). The participants gave their consent to participate in this study. A survey was sent via email to kindergarten to 12th-grade female teachers from randomly chosen school districts throughout the USA, primarily focusing on Michigan, Utah, Tennessee, Alabama, Florida, Georgia, New Hampshire, and South Carolina. Responses were received from teachers in more than 30 states throughout the USA. Within the email, a link to the survey was provided. Some paper copies of the survey were completed in both Tennessee and Utah by colleagues in those areas. Finally, various LISTSERV databases associated with teacher/education organizations were used to distribute links to this survey and obtain grade-level-specific information.
Only responses from female teachers were used in the analysis, because previous research has indicated a higher prevalence of voice disorders in females than males [4]. Given that most teachers are female and most of the survey respondents were female, focusing on female teachers in the analysis is a prudent focus. As such it was anticipated that results from this population would aid most in the understanding of vocal fatigue, its symptoms, and its development over time.
Methods: Survey
The online survey included 61 questions and was implemented by means of Qualtrics Survey software. To avoid overtaxing and high dropout rates, the length of the survey was set up to be completed in less than 15 min. The questions included the 19 items of the VFI and other vocal fatigue factors [20], as well as several questions inquiring about classroom capacity, class capacity (number of students), and teachers’ age. Regarding the classroom capacity, the teachers were asked to describe the classroom where they primarily teach by choosing from among the following options: small room (i.e., office, special needs room, 5–10 student capacity); medium room (i.e., general core classroom, art room, 16–35 capacity); large room (i.e., shop, music or performance room, 35+ capacity); and very large room (i.e., gymnasium, cafeteria, auditorium, outdoors).
Analysis
Poisson log-linear generalized linear models (GLMs) were used to test the associations between VFI (the three components of performance, pain, and recovery, as well as the total VFI) and primary classroom capacity (four levels: small, medium, large, very large) and age (considered as a continuous variable).
Deviance of the Poisson log-linear GLM was tested against degrees of freedom using χ2 distribution in order to detect and then account for potential overdispersion, which is common for this type of data [21]. The statistical significance of each parameter was then assessed with likelihood ratio-based χ2 statistics. A post hoc multiple comparison test was used after the GLM test to capture differences between classroom capacities to identify differences between groups. The analysis was conducted using the “R Project for Statistical Computing” software (R Core Team, 2014) implemented with the “AER” package (“dispersion test” function for overdispersion detection) and the dispmod package (“glm.poisson.disp” function for Poisson log-linear GLM overdispersion correction). The nested models were compared by χ2 test.
As the three sections of the survey were analyzed using a single summed vector method of analysis [10], if all of the questions of the VFI were not answered, the given subject was not included in the analysis.
Results
Of 570 female teacher respondents, 506 had fully completed their surveys at an average completion time of 12.5 min. Among those 506, 50 teachers reported that they were teaching in small classrooms, whereas 357 were teaching in medium ones, 88 in large ones, and 11 in very large ones. The average reported VFI scores for the individual scored factors (performance, pain, and recovery) were 16.7, 5.9, and 2.4, respectively. The average age of the respondents in this study was 42.9 years.
Using the VFI values suggested by Nanjundeswaran et al. [10], it was found that over 60% of the teachers who responded indicated that they were currently experiencing vocal fatigue. This prevalence of vocal fatigue is consistent with that found by Roy et al. [3], who stated that 58% of the teachers reported having issues with proper voice use or sound at the time they were surveyed.
The following three subsections provide the statistical results for each of the three factors categorized by the VFI. The distributions of the VFI scores associated with the three factors were most consistent with the Poisson distribution, according to the log-likelihood and plots of the fit of various possible distributions. Hence, the GLM was fit with the Poisson family of distributions and the log link function.
VFI Performance (Factor 1)
An overdispersion-corrected Poisson log-linear GLM was fit with the response variable of VFI performance (factor 1) (performance) and the covariates (1) classroom capacity and (2) age of teacher. The reference level for classroom capacity was “small.” The output of this model is reported in Table 1. The goodness-of-fit model was evaluated with the Hosmer-Lemeshow test, which assesses if the difference between observed data and predicted data is statistically significant (χ2 = –1.11, df = 8, p = 1).
As indicated in Table 1 and shown in Figure 1, VFI performance (factor 1) increased when the classroom capacity was bigger. The mean value scored by the teachers was 12.38 (standard error [SE] 0.93) in small classrooms, 17.02 (SE 0.42) in medium classrooms, 19.39 (SE 0.93) in large classrooms, and 22.55 (SE 2.25) in very large classrooms. As indicated by Tukey’s post hoc multiple comparisons, the responses to the questions in factor 1 were higher in larger classrooms and all the differences were statistically significant with the exception of the classroom pairs very large-medium and very large-large, due to the low number of respondents teaching in very large classrooms (medium-small: estimate = 0.32, z = 4.55, p < 0.001; large-small: estimate = 0.46, z = 5.47, p < 0.001; very large-small: estimate = 0.58, z = 3.88, p < 0.01; large-medium: estimate = 0.14, z = 2.57, p = 0.04; very large-medium: estimate = 0.26, z = 1.94, p = 0.18; very large-large: estimate = 0.13, z = 0.88, p = 0.79). The increased amount of variability in the results of teachers in very large classrooms is due in large part to having fewer responses from this group than from those teaching in smaller classrooms.
Regarding the effect of age on factor 1, the preferred model was the one with second-grade polynomic age terms. This means that the relation to factor 1 is not monotonic but there is an inflection point that is changing the direction of the relation between factor 1 and age. Figure 2 shows the aforementioned relationship, with the inflection point located between 40 and 45 years.
VFI Pain (Factor 2)
An overdispersion-corrected Poisson log-linear GLM was fit with the response variable from VFI pain (factor 2) (pain) and the covariates (1) classroom capacity and (2) age of teacher. The reference level for classroom capacity was “small.” The output of the model is reported in Table 2. The goodness-of-fit model was evaluated with the Hosmer-Lemeshow test (χ2 = –4.39, df = 8, p = 1).
As indicated in Table 2 and shown in Figure 3, the VFI pain (factor 2) score increased when the classroom capacity was bigger. The mean value scored by the teachers was 4.5 (SE 0.51) in small classrooms, 6.09 (SE 0.22) in medium classrooms, 7.06 (SE 0.47) in large classrooms, and 9.60 (SE 1.71) in very large classrooms. As indicated by Tukey’s post hoc multiple comparisons, VFI pain (factor 2) was higher in larger classrooms and all the differences were statistically significant with the exception of the pairs large-medium, very large-medium, and very large-large (medium-small: estimate = 0.30, z = 2.82, p = 0.02; large-small: estimate = 0.47, z = 3.82, p < 0.001; very large-small: estimate = 0.72, z = 3.24, p = 0.05; large-medium: estimate = 0.16, z = 2.13, p = 0.13; very large-medium: estimate = 0.41, z = 2.08, p = 0.14; very large-large: estimate = 0.25, z = 1.19, p = 0.60). The increased amount of variability in the results of teachers in very large classrooms is due in large part to having fewer responses from this group than from those teaching in smaller classrooms.
Regarding the effect of age on factor 2, the preferred model was the one with second-grade polynomic age terms. This means that the relation to this factor is not monotonic but there is an inflection point that is changing the direction of the relation between factor 2 and age. Figure 4 shows the aforementioned relationship, with the inflection point located between 40 and 45 years.
VFI Recovery (Factor 3)
An overdispersion-corrected Poisson log-linear GLM was fit with the response variable of VFI recovery (factor 3) (recovery) and the covariates (1) classroom capacity and (2) age of teacher. The reference level for classroom capacity was “small.” The output of the model is reported in Table 3. The goodness-of-fit model was evaluated with the Hosmer-Lemeshow test (χ2 = –17.61, df = 8, p = 1).
As indicated by the model output, the effect of classroom capacity on the responses to the factor 3 questions was not statistically significant, with the exception of the difference between small and very large classrooms. However, Tukey’s post hoc multiple comparisons showed that also the aforementioned difference was not statistically significant once the correction was applied.
Regarding the effect of age on factor 3, the preferred model was the one with second-grade polynomic age terms. This means that the relation between responses about vocal recovery is not monotonic but there is an inflection point that is changing the direction of the relation between factor 3 and age. Figure 5 shows the aforementioned relationship, with the inflection point located between 40 and 45 years.
Discussion
Teachers in higher-capacity (presumably larger-sized) classrooms reported increased VFI scores in all three factors (performance, pain, and recovery). This is consistent with previous research [22] and may provide a link between vocal fatigue and vocal effort measures. It may be that the increased effort associated with speaking in larger classrooms is creating an increase in the amount of vocal fatigue that teachers are reporting.
Specifically, the teachers reported vocal fatigue to be significantly elevated in the vocal performance factor of the VFI when compared to the lowest capacity classroom size (5–10 students). However, the elevated scores were not in a range high enough to be considered signifying vocal fatigue in that factor according to Nanjundeswaran et al. [10] (≥24 is considered “fatigued” in this factor). In factor 2 of the VFI, however, both teachers in large and very large classrooms averaged >7 out of a possible 20, indicating vocal fatigue in this factor (large: 7.06; very large: 9.6). This average score indicates a high amount of vocal fatigue as compared to nonoccupational voice users and vocally healthy individuals. Regarding factor 3, the female teachers reported a decreased ability to recover with rest when teaching in very-large-capacity classrooms compared to small-, medium-, and large-capacity classrooms (7 out of a possible 12 for this factor). These results generally indicate a significant effect of classroom capacity (number of students) on teachers reporting vocal fatigue, which corresponds with previous findings that room acoustics may affect teachers’ vocal health [23, 24].
With regard to the research question, it was predicted that VFI scores would have a positive correlation with primary classroom capacity and teachers’ age. It was hypothesized that those teachers who teach in larger-capacity classrooms would report higher amounts of vocal fatigue as there would be more noise in the classroom (with more students). Teachers would be obliged to use a louder voice to instruct in these large classrooms as their voices would need to carry throughout the entire room. Teachers instructing in more reverberant classrooms, which are also larger in capacity, experience higher amounts of vocal effort and a higher likelihood of voice problems [23, 11]. Therefore, the results of the current study may be used to raise awareness among teachers about the potential vocal risks of speaking for extended periods of time in a large classroom (especially one with poor acoustic properties).
Female teachers’ reports of elevated levels of vocal fatigue when teaching in a larger classroom may be due in part to their increased vocal demands in this setting. Despite the guidelines published by the ANSI regarding optimal classroom acoustics (ANSI S12.60-2002) mentioned above, the majority of school classrooms surveyed did not meet these requirements. For example, in an active classroom, sound levels may reach 50–60 dB [25], and if a signal-to-noise ratio of +15 dB is to be maintained [26], a teacher would be required to speak with a raised (66 dB(A)) to very loud (78 dB(A)) voice (following the definition of the ISO 9921) during the main part of the instruction period. The necessity of maintaining a sufficiently high voice level is likely a major reason why school teachers experience such elevated amounts of vocal fatigue.
As expected, the scores of the VFI associated with performance (general tiredness of voice) and pain (physical discomfort associated with voicing) were directly proportionate to the classroom capacity. However, the recovery score was not associated with the capacity of teachers’ primary classrooms. The reason for this may be that while the first two components of the VFI can be affected by environmental factors such as room capacity, recovery from vocal fatigue is more related to the physiology of the vocal mechanism. This is evidenced by the effect of age on vocal recovery.
While treated primarily as a control variable, it was assumed that older teachers would experience higher amounts of vocal fatigue due to physiological changes that occur with age [3]. However, in this study, the relationship between age and the VFI in all three factors appeared to be quadratic rather than linear. The highest values on the VFI were reported by teachers between 40 and 45 years of age, indicating the lowest levels of performance, the highest amounts of pain, and the longest recovery times. A possible explanation for this phenomenon could be that teachers younger than 40–45 years produce hormones which affect external tissue hydration [27] and make the vocal folds stiffer than in females older than 40–45 years. This quadratic relationship between teachers’ room capacity and their reported vocal pain, performance, and recovery may also be associated with teaching experience and the increased ability to adapt their voices to physiological and hormonal changes they experience over time. It is also conceivable that as teachers age, they are participating in fewer vocally demanding activities both during and after school. Additionally, older teachers with more tenure may be better respected and thus may be better able to manage classes. Teachers older than 60–65 years may also be scaling back on the number of classes they teach in a school day and thus are demanding less of their voice. Additionally, measurements of performance, pain, and recovery in female teachers may be affected by the average number of vocal fold collisions (related to their pitch). Due to this generally increased rate of collisions in the vocal folds during speech, females younger than 40–45 years may be more prone to poor vocal performance, increased pain, and increased recovery times.
Steps were taken to enhance the validity and authenticity of the survey responses. Only 11% (64/570) of the participants dropped out after completing less than half of the survey (these responses were not included in the analysis), indicating that most participants who began the survey were willing to complete it. To further verify the authenticity of the responses, most surveys were distributed by school administrators or to teachers’ email addresses obtained directly from school websites and LISTSERV databases. Finally, one question within the survey was an “attention filter” that identified those participants not paying attention to what they were answering if they filled in an answer other than the one instructed. This occurred rarely (<3%). Cultural differences and native languages spoken were not accounted for in this survey, and may have led to dropouts or fewer completed surveys.
Conclusion
Teachers in general, and female teachers specifically, have a higher instance of voice problems. The female teachers responding to our survey reported an elevated level of vocal fatigue, as quantified by the VFI. Additionally, with regard to the research question, a connection was found between vocal fatigue and teachers’ classroom size. Together, the vocal fatigue factors of vocal pain, performance, and recovery may have an effect on the vocal fatigue experienced by those who require the use of their voices to complete the day-to-day tasks of their jobs. Additionally, these factors may be affected by teachers’ primary room capacity.
Additionally, our research suggests that the female teacher-reported level of vocal fatigue increases from the age of 20 until the age of 40–45 years. After this time, it seems that age no longer has a negative effect on vocal fatigue. This may be an indication of teachers’ ability to adapt to the vocal demands of their work. It may also indicate a leveling off of biological changes occurring as women age.
Generally, these findings may be able to inform vocal health professionals, school administrators, and female teachers of the risks involved in teaching in larger- capacity classrooms. It may also assist in building an awareness of the need for taking better care of the voice, especially among teachers from the age of 20 to about 45 years, as this population reported higher risks for vocal fatigue.
Future work will include more detailed surveys and analyses of other risk factors, which could lead to additional insights and even modification of how and where teachers teach. Future research in this area will undoubtedly compare teachers’ objective measures of voice in addition to the scores on the VFI. Perceptual ratings by trained listeners should also be used as a comparison to teachers’ VFI scores and objective measures collected. Such work may give further insight into the amount of vocal fatigue experienced by teachers, as well as the notable subjective and objective differences in teachers’ voices teaching in various classroom capacities.
Acknowledgement
I would like to thank all the members of the Michigan State University Voice Biomechanics and Acoustics Laboratory and especially Alyssa Rollins, Allison Woodberg, Callan Gavigan, Lauren Glowski, and Sam John for their assistance in collecting teacher emails. Thanks to Drs. Chris Gaskill at the University of Monevallo (Alabama), Lynn Maxfield at the University of Utah, and Chaya Nanjundeswaren at East Tennessee State University for their help in collecting teacher responses to this survey. Thanks also to Dr. Simone Graetzer of the University of Liverpool for her contributions to this work.
The research reported in this publication was partially supported by the National Institute on Deafness and Other Communication Disorders of the National Institutes of Health under Award No. R01DC012315. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Disclosure Statement
The authors have no conflict of interest to declare.