What is Noise Sensitivity?

Abstract

Background:

Noise sensitivity moderates the association between environmental noise exposure and annoyance and health outcomes.

Methods:

In normally hearing adults, we measured noise sensitivity in three ways: using the noise sensitivity questionnaire, a 3-point self-rating, and the loudness discomfort level (LDL; mean reported discomfort level for tone bursts). We then presented recordings of a 15-second 80 dBLAeq aeroplane overflight and participants rated the annoyance and loudness they experienced.

Results:

The three measures of noise sensitivity were not well correlated with each other, and only the overall LDL was associated with the ratings of loudness and annoyance in response to the aeroplane sounds.

Conclusions:

This implies that our current measures of noise sensitivity may only capture parts of the underlying construct, and therefore underestimate effects due to it on the association between environmental noise and annoyance and health outcomes. We developed a theoretical model to describe the set of factors that may influence a person’s sensitivity to noise and propose that interaction between the systems described is the basis for noise sensitivity. This paradigm alters the focus of noise research from the annoyance caused by the sound, to the sensitization to noise that may occur as a result of the interplay of many factors. We hope that our model will allow research to explore the sensitizing factors for noise more easily and systematically.

INTRODUCTION

Environmental noise causes interference, distraction, and annoyance, and has direct and cumulative human health effects, which are numerous, persistent, and pervasive.[1] Annoyance due to noise is an important mediator in the relationship between noise, stress, and health.[2,3] People who are annoyed by noise experience negative emotions including fear and anger, accompanied by physiological arousal, which reinforces initial affective reactions, leading to negative health effects.[4]

The term “noise sensitivity” is used to refer to several concepts, which may depend on a single underlying dimension or may be the result of a more complicated set of processes. Noise sensitivity has been treated as a personality trait which is predictive of annoyance and health-related outcomes from noise in general.[5,6,7] In this usage, higher noise sensitivity is associated with subjective ratings of poorer health in those who are noise exposed but not in people without significant noise exposure.[6,7] It appears to be a personality trait in its own right, and is not explained by general affect.[8]

In early research on aircraft noise annoyance, sensitivity to noise was assessed in three different ways[9]: firstly, a set of questions (Guttman scale) that captured attitudes toward noise in general; secondly, annoyance and disturbance to activities caused by noise other than aircraft; and thirdly, a measure (called “neuroticism” by the author) captured the tendency to be annoyed by quieter sounds.[9] Independent associations between the first two measures and annoyance to aircraft noise were found, but “neuroticism” did not relate to annoyance with aeroplanes.[9] These findings suggest that the conceptualization of noise sensitivity may be contextual and might have multiple aspects.

The Weinstein noise sensitivity (WNS) scale, developed to measure subjective noise sensitivity, used a 6-point rating of 21 items regarding general attitudes to noise and emotional reactions to a variety of sounds.[10] The WNS has good internal reliability, test–retest reliability, and external validity,[11] and can be shortened to remove items that relate to measured noise exposure (Ldn) to ensure that it reflects specifically trait noise sensitivity.[12] Other shortened scales have also been developed for use in field studies.[13]

A multidimensional approach was used in the German-language questionnaire called Fragebogenzur Erfassung der individuellen Lδrmempfindlichkeit (LEF), across seven areas assessing everyday life, recreation, health, sleep, communication, work, and noise in general.[14,15] The noise sensitivity questionnaire (NoiSeQ) was based on adaptation and reformulation of the LEF and WNS questionnaires to assess noise sensitivity more systematically and to increase face validity. The NoiSeQ has 35 items, rated with a 5-point scale ranging from “strongly agree” to “strongly disagree.” Either an overall measure of noise sensitivity or separate noise sensitivity domains can be derived from subscales for work, leisure, habitation, communication, and sleep.[16] The NoiSeQ is effective for low to high noise sensitivity values, and is not influenced by sex or age, has good internal reliability, and has been used widely in field studies.[2,16,17]

Single-item ratings of noise sensitivity have been used in field research where longer questionnaires would be impractical.[18,19,20] It has been shown that single-item measures may underestimate noise sensitivity,[21] while also being slightly less strongly correlated with longer measures of noise sensitivity.[15] We have used a single 3-point (not noise sensitive, about average, highly noise sensitive) measure of noise sensitivity (3-NS) that appears to strike a useful balance between ease of responding and having sufficient validity in capturing noise sensitivity to produce meaningful results.[6,7] However, there is little research comparing single-item assessments against other noise sensitivity measures.

An alternative approach to assessing noise sensitivity is a direct, psychoacoustical approach conducted under laboratory conditions. Loudness discomfort levels (LDLs) are the levels of pure tones, across frequencies, at which a person reports discomfort to sounds.[22] They are widely used clinically and have been used for research purposes where they appear to relate to other measures of noise sensitivity.[23,24]

Noise sensitivity appears to be an important moderator of the association between environmental noise exposure and health effects in public health research, so it is important to understand its measurement properly. The first phase of the current research aimed to compare three measures: LDL, the NoiSeQ, and a 3-NS. If all three capture the same underlying construct, scores would be expected to correlate. On the other hand, if noise sensitivity depends on a process of interacting systems, correlations would be expected to be weaker.

Despite significant reductions in the noise level produced by aeroplanes generally, aircraft noise is rated as a highly annoying environmental noise source.[25] Interestingly, people report being more annoyed by a given level of aircraft noise exposure now than in the past, which may reflect sensitization to noise.[26,27,28] In addition, people who are more noise sensitive tend to experience more adverse health effects when exposed to noise from aircraft.[6] The second phase of the research investigated how well the three different measures of noise sensitivity predicted noise annoyance and/or the perceived loudness of the sounds of aeroplane overflights. As noise sensitivity is believed to moderate the response to noise, we hypothesized that the measures of noise sensitivity would predict annoyance and loudness outcomes.

METHODS

Participants

Thirty (11 males; 19 females) adults aged 21 to 67 (mean = 29.4, median = 24, standard deviation = 12.2) participated in the current study. Recruitment was through posters, social media, and electronic flyers, and most of the participants were university students or staff. Informed written consent was obtained. Participants were instructed to avoid the consumption of stimulants such as caffeine (4 hours), or alcohol (8 hours), or any sympathomimetic and/or anticholinergic drugs prior to testing. Hearing thresholds of all participants were no poorer than 25 dBHL (250–8000 Hz) bilaterally. Participants reported no history of cardiovascular or neurologic issues.

The research was approved by the University of Auckland Human Participants Ethics Committee (Ref: 18771).

Procedure

Participants were seated in a standard sound-attenuating chamber.[29] The dimensions of the booth were 2.2 m × 2.5 m. Ambient noise levels were compliant with the standard for maximum permissible ambient noise levels in an audiometric test room.[30] Sounds were delivered via Panasonic RP-HT160 Stereo Headphones.

Phase 1: Three measures of noise sensitivity

• (1)The NoiSeQ[14] was completed by each participant. The 35-item questionnaire asks about sensitivity to noise across five domains of life. Each participant was asked to imagine each situation and to indicate the extent of agreement with each statement on a 5-point (1–5) Likert scale, ranging from “strongly agree” to “strongly disagree,” for example, “I can do complicated work even while background music is playing.” The responses to 26 of the 35 NoiSeQ items were reverse coded so that the total score was higher if participants gave responses consistent with being more noise sensitive. The overall score was operationalized for each participant as the percentage of the maximum possible NoiSeQ score of 175 (35 × 5), so that values closer to 100 represent greater noise sensitivity. Scores for the five subscales of the NoiSeQ (leisure, work, habitation, communication, sleep) were computed following a parallel approach.
• (2)A single-item, 3-NS rating scale was used to classify the degree of noise sensitivity. Each participant was asked to indicate their own noise sensitivity on the scale: (1) low noise sensitivity, (2) average noise sensitivity, (3) high noise sensitivity.
• (3)The LDLs across the octave frequencies of 250 to 8000 Hz were obtained for both ears of participants using a standard technique.[31] The initial stimulus intensity was 50 dBHL, and frequencies were tested in the order: 1000, 2000, 4000, 8000, 500, 250 Hz. Pure tones were presented for approximately 2 seconds, with a 1-second interval between each presentation. A 2-dB step size was employed, until the participant indicated discomfort by pushing the response button. The physical and emotional state of the participant was carefully monitored throughout the testing procedure for possible signs of discomfort such as wincing, flinching, or frowning. Two ascending runs were completed per frequency, in which the higher presentation level was recorded and used for data analysis. When the discomfort level exceeded the limits of the audiometer, a value of 4 dB greater than the limiting value for the LDL response was recorded. A conservative safety limit was set at 106 dBHL and was not exceeded at any frequency. The responses across frequencies were highly correlated within each participant (Cronbach alpha = 0.98), so on this basis, an “overall LDL” was calculated by taking the arithmetic mean across both ears and all frequencies for each participant to create a total score.

Phase 2: Assessment of noise annoyance and perception of loudness

A large passenger jet aeroplane in a steady overflight was recorded with a sampling rate of 44,100 Hz, stereo channel selection, and a 16-bit depth on Adobe Creative Cloud, Audition® CC 2017 (v.10.1), Adobe Creative Cloud, and presented at a level of 80 dBLAeq. The 15-second “overflight” was preceded and followed by silence. Noise reduction techniques were applied and wind noise removed from the recording which was trimmed to 15 seconds. Onset and offset were smoothed using a spline curve. The frequency spectrum and relative levels were not altered from the original recording. Peak clipping was not apparent. Spectral analysis of the noise showed a dominance of middle frequency components (750–1500 Hz).

Rating of loudness and annoyance

Immediately after each aircraft overflight noise condition, participants provided responses on the perceived loudness and annoyance scales. Each participant experienced three repeats requiring separate loudness and annoyance ratings, and their final score was the mean of the three responses. We used different rating scales (a 9-point scale for loudness and an 11-point scale for annoyance) to discourage participants from indicating the same number for both scales without due consideration. The 9-point loudness scale ranged from (1) soft to (9) loud.[32] The 11-point numerical scale, developed by the International Commission on the Biological Effect of Noise, was used to assess annoyance; it ranged from (0) Not at all annoyed to (10) Extremely annoyed.[33]

Data treatment and analyses

Preliminary analyses were conducted using descriptive statistics to characterize the main measures. Correlation and ordinary least-squares multivariate linear regression were used in the main analyses. Where data met the assumptions of parametric tests (normality of residuals), these were preferred, but if assumptions were violated, nonparametric equivalents were used. One data point on the LDL measure was identified as an outlier, and that participant was included in preliminary descriptive statistics and correlations, then removed from all further analyses.

RESULTS

Phase 1

Participant ratings varied on all three measures of noise sensitivity. Responses to the 3-NS and summary scores on the NoiSeQ and overall LDL are presented in Table 1.

Table 1.

Description of participant ratings on the NoiSeQ and overall LDL; and counts (and percentages) responding on each category of the 3-NS

Measure	N	Minimum	Maximum	Mean	SD
NoiSeQ	30	37.71	86.86	63.64	13.38
Overall LDL	30	48.38	106.00	86.92	12.41
		Low	Average	High
3-NS	30	8 (27%)	15 (57%)	5 (17%)

SD, standard deviation; NoiSeQ, noise sensitivity questionnaire; LDL, loudness discomfort level; NS, noise sensitivity.

The minimum score on the LDL was unusually low at 48 dBHL, a finding suggesting a condition such as hyperacusis, or a failure to properly follow instructions.

Scores on the three measures of noise sensitivity were only weakly associated. The NoiSeQ and the overall LDL score correlated r = −0.260, P = 0.164 [Figure 1]. The 3-NS scale did not associate strongly with the other measures, with Spearman correlations of 0.320 (P = 0.084) with the NoiSeQ and −0.299 (P = 0.108) with the overall LDL [Figure 1].

Figure 1.

Score on the NoiSeQ as a function of overall LDL. The key indicates each participant’s score on the 3-point noise sensitivity scale. LDL, loudness discomfort level; NoiSeQ, noise sensitivity questionnaire.

After removal of the data point with an LDL <50 as an outlier, associations dropped further: NoiSeQ and overall LDL (r = −0.176, p=0.362); 3-NS and NoiSeQ (Spearman Rho = 0.168, P = 0.160); and 3-NS and overall LDL (Spearman Rho = −0.239, P = 0.211). As the unusual result appeared to inflate associations inappropriately, it was removed from further analyses.

Correlations between the NoiSeQ subscales, overall LDL score, and 3-NS were similarly low. The NoiSeQ subscales all correlated very highly with each other and the overall NoiSeQ score (r = 0.8–0.9), so further analyses were based only on the overall score.

Phase 2

Ratings of the loudness and annoyance experienced as a result of exposure to the 15-second 80 dB LAeq aeroplane overflight are presented in Table 2. Scores on the two measures were highly correlated (r = 0.834, P < 0.001) [Figure 2].

Table 2.

Description of the mean ratings of loudness and annoyance experienced after exposure to the sound of an 80 dB LAeq aeroplane overflight

Measure	N	Minimum	Maximum	Mean	SD
Loudness	29	3.00	9.00	6.74	1.50
Annoyance	29	2.67	9.67	6.10	2.02

SD, standard deviation.

Figure 2.

Annoyance rating (on a scale of 0–10 where higher values indicate greater annoyance) plotted as a function of loudness rating (on a scale of 1–9 where higher values indicate greater loudness) in response to a recorded aeroplane overflight presented at 80 dB LAeq for 15 seconds.

In bivariate analyses, overall LDL predicted the annoyance [Figure 3] and loudness [Figure 4] of aeroplane sound significantly, whereas the other two measures did not reach statistical significance, though the direction of effects from all of the measures was that more noise-sensitive people tended to rate the sound as louder and more annoying [Table 3].

Figure 3.

Perceived loudness on a rating scale from 1 to 9 of a 15-second, 80 dB LAeq aeroplane overflight plotted against overall loudness discomfort level (overall LDL).

Figure 4.

Annoyance rating from 0 to 10 of a 15-second, 80 dB LAeq aeroplane overflight plotted against overall loudness discomfort level (overall LDL).

Table 3.

Associations between measures of noise sensitivity and rated loudness and annoyance to an 80 dB LAeq aeroplane overflight

	Loudness			Annoyance
	β-coefficient	P-value	Adjusted R2	β-coefficient	P-value	Adjusted R2
Bivariate associations
NoiSeQ	0.180	0.350	-0.003	0.337	0.074	0.081
3-NS	0.323	0.087	0.071	0.248	0.194	0.027
Overall LDL	−0.433	0.019	0.158	−0.413	0.026	0.140
Combined model			0.153			0.164
NoiSeQ	0.067	0.714		0.252	0.170
3-NS	0.221	0.236		0.112	0.542
Overall LDL	−0.369	0.051		−0.342	0.068

The upper half of the table shows bivariate associations and the lower half shows the associations in multivariable models with all three measures of noise sensitivity included. NoiSeQ, noise sensitivity questionnaire; LDL, loudness discomfort level; NS, noise sensitivity.

Combining the three noise-sensitivity measures in a single linear regression model did not improve the model (as indicated by adjusted R²) over that provided by overall LDL alone [Table 3]. Furthermore, the significant effects observed in the bivariate models for overall LDL were reduced, implying that the variance in loudness and annoyance explained by the three measures of noise sensitivity was shared.

DISCUSSION

In the first phase, three measures of noise sensitivity were made in a group of normally hearing adults: a questionnaire (NoiSeQ), a single 3-point rating scale (3-NS), and a behavioral assessment based on reported loudness discomfort to tonal stimuli played at gradually increasing levels (overall LDL). These measures were not well correlated. In the second phase, participants listened to recordings of an aeroplane overflight presented for 15 seconds at 80 dBLAeq, and rated their experience of loudness and annoyance associated with the sound. Only overall LDL predicted loudness and annoyance ratings significantly. In models with all three measures of noise sensitivity as predictors, the overall prediction did not improve, implying that the variance in loudness and annoyance explained by the three measures of noise sensitivity was shared.

The associations between the measures of noise sensitivity used here were clearly weaker than in previous research into the association between different measures of noise sensitivity.[15] A difference between these studies and the current is that we used three very different approaches to measure noise sensitivity. The NoiSeQ is a short but quite detailed questionnaire that explores noise sensitivity both overall and in different domains; it might be observed to parallel the questionnaire measures used by Zimmer and Ellermeier. By contrast, the 3-NS takes an extremely simplistic approach by asking people overall and in general to rate their own sensitivity to noise compared to others, without specifying the location/domain, and using only 3 points. The LDL approach seems to capture a concept that is not asked about in the NoiSeQ: intolerance to sound level. The LDL gives a precisely measured set of discomfort levels, which were strongly intercorrelated, and the combined variable “overall LDL” might have had more precision and thus variability that enabled associations to be detected with the outcome measures. Despite the different measurement techniques, based on their use in previously published research, we regarded all three methods as valid ways to assess participants’ noise sensitivity.

There were some interesting trends apparent in Figure 1 that were not revealed in the statistics. This may be because the apparent patterns were not consistent across the full range and thus not captured well by statistical modeling. The participants who indicated that they had high noise sensitivity on the 3-NS (green squares) tended to score highly on the NoiSeQ, though their scores on the LDL were very variable. On the other hand, participants who indicated that they had low noise sensitivity on the 3-NS (red triangles) tended to have high LDL scores, though their NoiSeQ scores were highly variable. This may mean that our measures of noise sensitivity are each capturing parts of one or multiple underlying components.

It has long been understood that noise sensitivity, though often regarded as a personality trait, must be situational and the NoiSeQ aims to capture that aspect of it by asking questions about noise sensitivity in different domains. However, the fact that the overall reliability coefficient for the NoiSeQ is very high (>0.9), suggests that the domains of noise sensitivity it captures are related in the minds of respondents, so that if a person is sensitive to noise in one domain, they tend to be so in others as well. The weak associations observed between the three measures in the current research may indicate that there are other aspects of noise sensitivity that are not captured in the NoiSeQ.

Another area of research that may help to inform thinking about noise sensitivity is the soundscape; the perception and response to the acoustic environment.[34] One aspect of this is the expectation held by people in an environment, where sounds that are expected and observed as appropriate for that environment may be tolerated better than sounds that are observed as breaking the accepted acoustic range.[35]

Noise sensitivity is often defined as a single psychological trait, and on this basis, it is considered possible to measure it using a single rating or scale. We propose that this conceptualization as a simple and singular concept does not capture well its different aspects.[36] There are many possible interpretations of noise sensitivity, for example: loudness intolerance, ease of distraction from tasks by noise, likelihood of sleep disturbance, being upset by loud noises, being irritated by quiet noises, having difficulty hearing in background noise, having a negative attitude toward sources of noise, inability to identify other sounds clearly due to masking effects. Thus, when a person is asked to rate their own noise sensitivity, they may consider just one, some, or all of these aspects. Moreover, there may be a degree of “priming” depending on the situation they are in, especially, if they have not previously considered their sensitivity to noise per se, so may weigh each component differently depending on the circumstances. Based on our understanding of the processes related to noise sensitivity, we propose that noise sensitivity is a process and that it occurs according to the system described in Figure 5.

Figure 5.

Process diagram of a system model of noise sensitivity. A sound has characteristics that may be detected by the auditory system. If detected, the percept of the sound is interpreted cortically for meaning, while in parallel, the information passes into the limbic system where it can contribute to physiological arousal, affect, and wakefulness. Depending upon a person’s state, situation, and what they are doing, combined with their psychological traits and their attitude to the source of the sound, they may interpret the sound as “noise.” In other words, noise sensitivity is not merely a psychological trait, but rather the result of a series of variables and processes that combine to produce it.

The components of the system model of noise sensitivity [Figure 5] are as follows:

• (1)Sound source: The entity that generates the sound or is responsible for its generation. For example, a person, a company, a road, or a vehicle. These examples demonstrate the varying and subjective nature of the perceived source of a sound.
• (2)Sound quantities: Basic physical/acoustical aspects of the sound which are measurable. For example, level, frequency, and spectral energy. As we propose considering noise sensitivity as a system, these aspects of sound should be considered as part of the causal pathway for it. This argument goes against the classical view that sounds of higher level are naturally more annoying, and proposes a paradigm where the level is one factor that influences a person’s sensitivity to a sound, as part of the system as a whole.
• (3)Sound qualities: The aspects of the sound that are objectively present but which are not captured in time/frequency-weighted averages. For example, fluctuations over time, structure, and phase relationships with other sounds.
• (4)Auditory system: The ears and brain-auditory pathways to the primary auditory cortex. Previous research has shown differences in the early cortical processing of noisier sounds in nonnoise-sensitive people, implying that behavioral noise sensitivity may be influenced by early cortical processing.[37] There have also been differences observed in the gray matter volumes of structures associated with auditory processing according to noise sensitivity.[38]
• (5)Sound percept: The perceptual experience of the qualities of the sound itself, without associated meanings; the aspects of sound processed in the primarily auditory areas of the brain: the perceived pitch, loudness, and location of the sound.
• (6)Meaning: The interpretation of sound into meaningful information. For example, speech or music.
• (7)Limbic system: The structures in the brainstem that mediate emotional experiences and which link these to physiology. For example, the amygdala, hypothalamus, reticular activating system.
• (8)State: Psychophysiological state, including emotions, etc. This may combine pre-existing states (e.g., anxiety, calm, intoxication) with the influences of the sound described above.
• (9)Arousal: physiological arousal via the autonomic nervous system.
• (10)Affect: Affective responses to sounds caused directly by the sound and mediated via the limbic system LeDoux’s “low road,”[39] as opposed to cortically mediated responses to the sound.
• (11)Wakefulness: Either the state of full wakefulness or the transition from deeper to lighter sleep stages potentially resulting from sound exposure.
• (12)psychological traits: Stable and lasting aspects of personality. For example, openness, conscientiousness, extraversion, agreeableness, neuroticism (OCEAN), and the personality trait of “noise sensitivity” as has been described elsewhere. Possibly this may include noise reactivity (the tendency to complain about noise).
• (13)Attitude to the sound source: The listener’s pre-existing attitudes toward the perceived source of the sound. For example, a group of friends, a company that employs them, a neighbor with whom there is a dispute. This has long been known to be related to noise sensitivity.[23,40] In some earlier research, this was captured as a set of attitudes including fear of danger, belief the noise could be prevented, belief about the importance of the noise source, and annoyance with nonnoise impacts of the source,[40] but this type of information is rarely collected.
• (14)Situation: where the person is located. For example, home, work, on holiday. This aspect of noise sensitivity is captured by the NoiSeQ to some extent. The situation may bring with it expectations as has been shown in research on the soundscape.[35]
• (15)Behavior: what the person is doing or wants to do. For example, conversing, sleeping, carrying out a task.
• (16)“The sound is noise”: captures the experience of the sound as noise: that is, distracting, annoying, uncomfortable, awakening, interfering, etc. Importantly, the decision that a given sound is noise indicates that it has been appraised negatively. The subsequent behavioral reaction to the presence of noise is not a matter of noise sensitivity, but of noise reactivity.

We propose, therefore, that noise sensitivity is a process involving the interplay of the factors described and shown in Figure 5. Some of the linking arrows are bidirectional; for example, heightened sensitivity to sound can occur with limbic system activation, and because the interpretation placed on sound can both influence and be influenced by limbic system activity.[39] We should also make it clear that this process does not occur in a vacuum: intrinsic factors such as ageing and other exposures influence the ability of the auditory system to detect stimuli, a prerequisite of the later components of the process. Other factors (e.g., sex and gender) may also influence the degree of hearing loss, and other areas of the process such as limbic system activity. Furthermore, broader cognitive functioning and mental health may be expected to influence many areas of the process.

Limitations

The use of aeroplane sound may not generalize to other types of sound exposure, and replication of the findings in aircraft and in other types of sound (e.g., road traffic) will be important. Furthermore, the order of the three different measures of noise sensitivity was fixed, with the NoiSeQ first, then the 3-NS, then the LDL. This might possibly have had an influence in how responses were captured. The system model of noise sensitivity was created to attempt to capture the key areas that may influence the decision to regard a stimulus as noise; however, it may not be complete, and further research will be important to test and expand it.

Summary

Implicit in much previous research, by ourselves and others, into the impact of noise was the treatment of noise as an environmental phenomenon in its own right. We then considered noise sensitivity as a unidimensional moderator: for a given noise exposure, people who were highly noise sensitive would experience more annoyance and health effects than would others. Now, we suggest that sounds should not be predetermined as “noise,” and propose that any sound exposure may potentially be interpreted as noise according to the complex process, depicted in Figure 5. “Noise sensitivity” in this definition is the process by which a given sound may be experienced as noise by someone exposed to it. This approach avoids applying a subjective value judgment to an environmental stimulus as if it were an objective assessment, and presents noise sensitivity as more than simply a trait.

CONCLUSION

In this study, we have investigated noise sensitivity in three ways and compared these measures to the experience of annoyance and loudness associated with an aeroplane overflight. The measures of noise sensitivity were not well correlated, and though they predicted annoyance/loudness, they did not do so strongly. In response to this, we identified the need for a more complex understanding of environmental noise sensitivity that takes into account the system of interacting factors that cause a person to be sensitive to sound. We have proposed a process model for noise sensitivity that will allow future research to focus on specific aspects of the system while also considering the other aspects contributing to noise sensitivity.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.