A Scoping Review of the Effect of Classroom Acoustic Conditions on Primary School Children’s Numeracy Performance and Listening Comprehension

Abstract

Obtaining adequate numeracy skills and listening comprehension skills at primary school are vital for children’s future success. However, classrooms are often noisy and reverberant which may interfere with learning these skills. Two scoping reviews were conducted to synthesise research assessing the effect of different classroom acoustic conditions on (1) children’s numeracy performance and (2) children’s listening comprehension and to identify areas for future research. The PRISMA-ScR protocol was used for these scoping reviews. A comprehensive search of four online databases was conducted in September 2021 using the search term classroom AND (noise OR reverberation OR acoustics) AND (numeracy OR math* OR arithmetic) for the first scoping review, and in May 2022 using the search term classroom AND (acoustic* OR noise OR reverb*) AND ("listening comprehension" OR "auditory comprehension" OR "spoken language comprehension" OR "speech comprehension”) for the second scoping review. The effect of the acoustic conditions on children’s numeracy was varied with most studies showing a negative or no effect of noise, but two showed a positive effect. Therefore, future research is needed to better understand the effect of different classroom acoustic conditions on children’s numeracy performance. For listening comprehension overall, signal-to-noise ratios below + 10 dB mostly had a negative effect on children’s listening comprehension compared to quiet conditions; however, variables such as the noise type, signal-to-noise ratio tested, the listening comprehension domain examined, the population studied, and the voice used for the stimuli affected this. Future research avenues to better understand these effects are proposed.

1 Introduction

The acoustic conditions of a classroom such as noise and reverberation can impact on children’s learning [1]. Specifically, a recent review showed that noise affects children’s literacy skills of reading, writing, and spelling [2]. The present review will focus on the effects of different classroom acoustic conditions on (1) children’s numeracy performance and (2) children’s listening comprehension.

Noise may affect children’s performance in numeracy and their listening comprehension. Noise sources can be external to the classroom, such as traffic, aircraft, construction, and weather noise. Internal noise sources can also be present including building and service noise such as noise from heating, ventilation, and air-conditioning systems and equipment noise, and noise from the children, such as speech and movement. Noise can be exacerbated by the reverberation in a room. Reverberation can be quantified by the reverberation time of the room, which is the time it takes for a sound to decay by 60 dB. Longer reverberation times mean the sound is more prolonged in the room. Historically, most studies have examined the effects of external noise on children’s performance; however, more recently, studies have been examining the effect of noise from the children within the classroom [3]. When examining the effect of noise on different processes, noise exposure can be classified as either chronic, i.e. noise exposure over a long period of time, or acute, i.e. noise exposure over a short period of time. Studies on chronic noise exposure usually examine real noise naturally occurring in the classroom and hence have ecological validity but less control. Studies on acute noise exposure usually have recorded noise presented in experimental conditions so allow for greater control but can be less ecologically valid.

Noise has been found to affect children’s performance on both auditory and non-auditory tasks. Regarding non-auditory tasks, noise can have a negative effect on children’s visual short-term memory, reading, and writing performance (see [4, 5] and [2] for reviews). Poorer attention control in children compared to adults can result in this poorer performance (see [4] for a review). Therefore, noise may affect children’s numeracy skills for mathematical questions that are presented in the visual domain.

For auditory tasks, noise impairs children’s speech perception (see [4] for a review). To succeed at school, it is vital that children are able to listen and comprehend their teacher’s and classmates’ speech. Factors which affect listening can be classified into listener-external and listener-internal factors [6]. Listener-external factors include the target speaker, interfering noise sources, spatial configuration, and the acoustic characteristics of the room. Listener-internal factors include the child’s auditory and cognitive processing (which could be affected by having a young, underdeveloped auditory system) and their personal state. Children do not reach adult-like performance on speech perception tasks in noise and reverberation until the late teenage years [7]. Children’s poorer auditory selective attention skills compared to adults can be attributed to them experiencing greater distraction from noise (see [4] for a review).

It has long been established that poor classroom acoustic conditions affect children’s speech perception [8], but what about the effects on listening comprehension? According to Kiessling et al. [9] listening “is the process of hearing with intention and attention” (p. 93). Comprehending, however, goes beyond this. Comprehending “is the reception of information, meaning, or intent” [9] (p. 93). Schiller et al. [10] have developed the Speech Processing under Acoustic DEgradations (SPADE) framework which proposes three processing dimensions (speech perception, listening comprehension, and auditory working memory) to determine the effect of noise on children’s ability to “(a) auditorily perceive what is being said, (b) understand (comprehend) the content of a verbal message, and (c) memorize what they have been told” (p. 171). The authors note that listening comprehension requires semantic and syntactic integration, requires top-down processing, and is influenced by prior knowledge and language abilities. Listening comprehension tasks involve longer speech stimuli than speech perception tasks and rather than specifically assessing speech decoding abilities in noise, they assess the effect of increased listening effort in noise on how well children understand what has been said [10]. Therefore, noise may affect children’s numeracy skills for mathematical questions that are presented in the auditory domain and affect their listening comprehension more generally.

There are several recommendations about what the acoustic conditions in a classroom should be with these mainly derived to ensure adequate perception of speech [11, 12]. Because of the adverse effects on speech intelligibility, many countries have room acoustic standards with recommended maximum unoccupied noise levels and reverberation times for classrooms. For example, the Australian/New Zealand Standard AS/NZS2107:2016 [12] recommends that unoccupied noise levels should not exceed 35–45 dBA and that reverberation times for standard classrooms should be around 0.5 s (though this does vary depending on the size of the classroom). For occupied classrooms, it is recommended in the research literature that the noise levels should not exceed 50 dBA to allow for a + 15 dB SNR given that a typical voice is around 65 dBA [8, 11, 13]. However, these conditions are often not met [11]. Schools may be built in high traffic areas or under flight paths and/or have reverberation times above the recommended 0.5 s (see [11] for a review). Additionally, modern teaching methods that focus on group work are becoming more common making up around 50% of teaching time [14, 15]. These activities have higher noise levels than lecture-style or independent learning [16, 17]. Furthermore, open plan innovative learning environments are growing in popularity [15] and have higher intrusive noise levels coming from the adjacent classes without walls between them [17].

Therefore, it is important to understand the effect that different classroom acoustic conditions have on children’s numeracy performance for tasks presented in the visual and auditory domains, and on their listening comprehension more generally. To achieve this understanding, two scoping reviews were carried out—one investigating the effect of classroom acoustic conditions on children’s numeracy performance and one examining the effect on listening comprehension. A scoping review as outlined by Munn et al. [18] aims “to identify the types of available evidence in a given field; to clarify key concepts/definitions in the literature; to examine how research is conducted on a certain topic or field; to identify key characteristics or factors related to a concept; as a precursor to a systematic review; and to identify and analyse knowledge gaps” (p. 2). In contrast, a systematic review has a focus on informing practice and policy which was not the purpose of this paper [18].

2 The Effect of Classroom Acoustic Conditions on Primary School Children’s Numeracy Performance

According to the Australian Curriculum, Assessment and Reporting Authority, “numeracy encompasses the knowledge, skills, behaviours and dispositions that students need to use mathematics in a wide range of situations. It involves students recognising and understanding the role of mathematics in the world and having the dispositions and capacities to use mathematical knowledge and skills purposefully” [19]. Poor numeracy skills can reduce a person’s employment opportunities and limit their job progression opportunities [20]. Employers in a wide range of disciplines highly regard graduates’ numeracy skills and often use numeracy tests as part of their recruitment process [21]. Therefore, competence in numeracy from the early years is of vital importance for a child’s future success.

The aim of this first scoping review was to synthesise and systematically map research that has assessed the effect of different classroom acoustic conditions on primary school children’s numeracy performance as well as to identify gaps to inform future research. The following research question was formulated: What is known from the literature about the effect of classroom acoustic conditions on primary school children’s numeracy performance?

3 Method

3.1 Protocol

The Preferred Reporting Items for Systematic reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR) [22] was the protocol used for this scoping review. The PRISMA extension for scoping reviews website can be found here: http://www.prisma-statement.org/Extensions/ScopingReviews.

3.2 Eligibility Criteria

The peer-reviewed papers had to meet the following criteria to be included in the review: (1) conduct a study on the effect of classroom acoustics (i.e. noise or reverberation) on children’s numeracy performance via a test (i.e. not a questionnaire) in either the field or the laboratory, (2) be conducted with primary school children (i.e. include children aged 5–12 years—papers that included slightly older children were also included as long as they included a group of children within this age range), and (3) have the full text in English available.

3.3 Information Sources

To identify potentially relevant documents, the following bibliographic databases were searched: ERIC, PubMed, Scopus, and Web of Science. The final search results were exported into.csv files where duplicates were removed.

3.4 Search

A comprehensive search of the online databases was conducted on 3 September 2021 to identify the effects of classroom acoustics on children’s numeracy performance. The search term was classroom AND (noise OR reverberation OR acoustics) AND (numeracy OR math* OR arithmetic). No publication date restrictions were applied.

3.5 Selection of Sources of Evidence

All publications identified in the searches were evaluated by the titles, and then abstracts and full text when needed for potentially relevant publications.

3.6 Data Charting Process

Data charting refers to how relevant information from the papers was extracted. Data from eligible studies were charted to capture the relevant information on key study characteristics of the effect of classroom acoustic conditions on children’s numeracy performance.

3.7 Data Items

Data were abstracted on the following characteristics: child populations that have been studied, the types of acoustic conditions that have been assessed, the types of measures used to assess numeracy performance, and the effect of the acoustic conditions on children’s numeracy performance.

3.8 Synthesis of Results

Studies were grouped by the acoustic conditions explored and summarised according to the effect of the acoustic conditions on children’s numeracy performance.

4 Results

4.1 Selection of Sources of Evidence

The search and selection process of the studies included in the review is shown in Fig. 1. A total of 227 papers (170 after removing duplicates) were returned in the searches. These were vetted for relevance via reading the title, abstract, and when needed for clarification, the full text. Nine papers were deemed relevant for the review. The references of these papers were checked and an additional six journal articles fitted the review criteria bringing the total number of papers to be reviewed to 15.

Fig. 1

Database search results for the numeracy performance scoping review

4.2 Characteristics and Results of Sources of Evidence

The general information for the 15 papers included in the review is shown in Table 1. The range of publication years was from 1972 to 2021, with the majority of papers being published since 2002 (see Fig. 2). The following sections describe the population studied, the acoustic conditions investigated, the measures and methods used to assess numeracy, and the outcomes of the papers.

Table 1 General information for the 15 papers included in the review and effect on children’s numeracy outcomes

Fig. 2

Publication years for the 15 papers reviewed in the numeracy performance scoping review

4.3 Population

All studies assessed the range of children found in mainstream classroom except for one study which specifically assessed the effect of noise on school performance in hyperactive children [23]. Johansson [24], however, did analyse the results of the study for children with high versus low intelligence as measured by a standardised general intelligence test [25]. The age range of children assessed was from second grade to eighth grade, i.e. 7- to 13-year-olds. (Note that the studies involving seventh and eighth grade children were included even though this was outside of primary school age group as they also included sixth graders.)

4.4 Acoustic Exposure

The acoustic exposure investigated in the reviewed papers can be split into two categories: chronic noise exposure and acute exposure.

Seven studies assessed the effect of chronic noise exposure on children’s mathematics performance. The source of chronic noise exposure was external noise such as aircraft noise [26, 27] or traffic noise [28], internal building and service noise such as noise from heating and cooling systems [29, 30], or internal occupied noise such as talking and movement noise from the children in the classroom [31, 32]. The details of the acoustic conditions can be found in Table 1.

Seven studies assessed the immediate effect of acute noise played via loudspeakers or over headphones on children’s mathematical performance [23, 24, 33,34,35,36,37]. The types of acute noise were intelligible speech [33], irrelevant speech [35, 36], babble [34], babble and environmental noise [34], music [33], music and speech [33], intelligible classroom noise [23], irrelevant classroom noise [36, 37], octave band noise [24], and road traffic noise [35, 37]. One study assessed the effect of acute noise played through loudspeakers in classrooms with different reverberation times [38]. The details of the acoustic conditions can be found in Table 1.

4.5 Measures and Methods

A range of tests were used to assess children’s mathematics performance (see Table 1) including standardised school tests and tests created specifically for the study. There were also a range of test presentations including written/visual presentation, auditory presentation, numerical equations, and problems in word form. Outcome measures included accuracy of solving the mathematical problem and speed of response.

Six studies used the results of standardised school tests to assess the impact of chronic noise exposure on children’s mathematical performance. Cohen et al. [26] used the California Test of Basic Skills tests scores gathered from the schools. Haines et al. [27] used the children’s United Kingdom National Standardised Assessment Test mathematics scores. This assessment was also used by Shield and Dockrell [31] and Shield and Dockrell [32]. Ronsse and Wang [29] used the children’s mathematics results from the Iowa Test of Basic Skills. Ronsse and Wang [30] used the children’s mathematics results from the Terra Nova tests.

Nine studies created mathematics tests specifically for the study. Kassinove [33] assessed children’s arithmetic calculation abilities with one group assigned to easy problems and the other group assigned to difficult problems. For the third grade children, the easy problems group were asked to add two one-digit numbers with no carrying required. For the difficult problems group, the children were asked to add two two-digit numbers that required carrying. For the sixth grade children, the easy problems group were asked to solve division problems with a two-digit number divided by a one-digit number. The difficult problems group were asked to divide a four-digit number by a two-digit number. Children were assessed on their response accuracy and how many problems they solved in 45 min.

Zentall and Shaw [23] assessed children’s performance on mathematics problems but no additional details about the form of the mathematics problems were reported. Johansson [24] assessed children’s performance on 406 multiplication items of two digits by one. The children were given 25 min to complete as many equations as they could and were scored on their speed and accuracy. Dockrell and Shield [34] assessed children’s arithmetic skills via a paper and pencil arithmetic test that the children worked through at their own speed. The test involved basic computations but no verbal component.

Ljung et al. [35] assessed children’s mathematical performance on arithmetical and geometrical problems, and a mathematical reasoning test. For the arithmetical tasks, the child answered three division problems and three multiplication problems. For the geometrical problems, the child was asked two questions on naming points in a coordinate system, four questions demonstrating understanding of the relationship between fractional expressions and areas of figures, two questions on understanding the relationship between distance and numerical expressions, and two questions measuring distances. The mathematical reasoning test comprised problems expressed in words with a numerical solution.

Papanikolaou et al. [28] assessed children’s problem solving skills and arithmetic calculations. No other details of the types of problems were provided. Meinhardt-Injac et al. [36] assessed children’s mathematical performance via 126 simple mathematical equations requiring addition or subtraction of two numbers. The children were asked whether a mathematical equation was correct or incorrect with accuracy and reaction time being recorded.

Caviola et al. [37] and Prodi and Visentin [38] assessed children’s mathematics performance using 28 two-digit additions and subtractions with two levels of difficulty. Low difficulty equations did not involve carrying or borrowing, whereas high difficulty equations did. The child listened to an audio recording of the problem and was asked to choose the correct response on their tablet from three multiple-choice answers (the correct answer, the correct answer plus or minus two, and the correct answer plus or minus 10). The children had up to 20 s to provide a response. Accuracy and response times were recorded for each problem.

4.6 Outcomes

An overview of whether increased noise/reverberation resulted in a negative effect, no effect, or positive effect on children’s mathematical performance can be found in Table 1. A summary of the findings categorised by the acoustic exposure studied is found in Appendix A.

Figure 3 graphically presents the effect of chronic noise exposure (negative or no effect) on children’s mathematical performance for children in different years collated from the reviewed papers. Papers could have more than one result depending on the grade, but also on the types of noise or schools studied. For example, the study by Cohen et al. [26] was scored twice as there was a negative effect on mathematics performance for high noise schools compared to schools with high noise but noise abatement, but no effect for high versus low noise schools. As shown in Fig. 3, there was a mix of negative effects and no effects of noise on children’s mathematical performance. No studies showed positive effects of noise. Due to the small number of studies, a clear trend for the effect of noise on children of different ages could not be determined.

Fig. 3

Effect of chronic noise exposure (negative, no effect) on children’s mathematical performance for children in different years collated from reviewed papers with references in square brackets

Figure 4 graphically presents the effect of chronic noise exposure (negative or no effect) on children’s mathematical performance for different noise types. Traffic noise was shown to have a negative effect in both of the studies exploring this noise source. Heating and cooling system noise at the levels studied did not appear to affect children’s mathematical performance. Conclusive findings could not be drawn for aircraft noise or classroom noise due to the mixed results and small number of studies for each type of noise.

Fig. 4

Effect of chronic noise exposure (negative, no effect) on children’s mathematical performance for different noise types collated from reviewed papers with references in square brackets

Figure 5 shows the noise levels examined in the studies and their effect (negative or no effect) on mathematical performance. Higher noise levels around the 70–80 dB mark tended to result in poorer mathematical performance. Shield and Dockrell [32], however, show a trend of poorer performance from 30 dB upwards with it steepening from 60 dB and above so keeping noise levels below this is even more ideal. Note that Cohen et al. [26] were not included in this figure as the noise levels reported were peak, not average.

Fig. 5

Effect of chronic noise exposure (negative, no effect) on children’s mathematical performance for different noise levels compared to the reference condition collated from reviewed papers. Lines represent when a range of levels were studied

Figure 6 graphically presents the effect of acute noise exposure (negative, no effect, or positive) on children’s mathematical performance for children in different years (including children with special educational needs (SEN) in hashed colours) collated from reviewed papers. Again, papers could have more than one result depending on the grade, but also on the types of noise studied. The results were mixed with some studies showing negative effects, some showing no effects, and some showing positive effects, so no trends for the data could be drawn.

Fig. 6

Effect of acute noise exposure (negative, no effect, positive) on children’s mathematical performance for children in different years (including children with special educational needs (SEN)) collated from reviewed papers with references in square brackets

Figure 7 graphically presents the effect of chronic noise exposure (negative, no effect, or positive) on children’s mathematical performance for different noise types (including children with special educational needs (SEN) in hashed colours). It appears that acute exposure to traffic noise does not affect children’s mathematical performance at the level studied. This finding is interesting given that a negative effect of chronic road traffic noise was found. This suggests that while acute exposure to road traffic noise may not affect children’s mathematical performance initially, long-term, chronic exposure may have a negative effect. Therefore, it is still advantageous to ensure that schools are built away from main roads if possible or are built with sound insulation. Otherwise, however, conclusive findings could not be drawn due to the mixed results and small number of studies for each type of noise.

Fig. 7

Effect of acute noise exposure (negative, no effect, positive) on children’s mathematical performance for different noise types (including children with special educational needs (SEN)) collated from reviewed papers with references in square brackets

Figure 8 shows the noise levels examined in the studies and their effect (negative, no effect, or positive) on mathematical performance. The noise levels studied were concentrated around 50–60 dB, but the effect of the noise was mixed so clear conclusions about the effect of different noise levels on children’s mathematical performance cannot be drawn. (Note that when the condition was silence, 25 dB was chosen to mark the reference condition.) However, given that several studies did report negative effects of noise on children’s mathematical performance, it would be recommended to keep noise levels in classrooms as low as they possible.

Fig. 8

Effect of acute noise exposure (negative, no effect, positive) on children’s mathematical performance for different noise levels compared to the reference condition collated from reviewed papers. Note that when the condition was silence, 25 dB was chosen to mark the reference condition. Lines represent when a range of levels were studied

For the two studies that did report positive effects of acute noise on children’s mathematical performance, there are some important points to note. Zentall and Shaw [23] only had a difference of 5 dB between the two noise levels (64 dB and 69 dB) and did not compare this to a quiet conditions where a negative effect of the higher noise level may have been found. The positive effect of noise in the Johansson [24] was specifically for those children with high intelligence (whereas those with lower intelligence performed worse in noise). These specific children may have found the noise to boost their stimulation levels to an optimal level rather than overstimulate them as may have been the case for the lower intelligence group. Therefore, it is important to note that these positive results were for very specific studies.

Figure 9 shows the effect of noise or reverberation exposure (negative, no effect, positive) on children’s mathematical performance when the task was presented in different modalities (visual, auditory, not specified) for the 15 reviewed papers. Mixed results are seen. It is important to note, however, that for the auditory domain, children at the top end of the age range included in the review were assessed (11–13 years) and negative effects of higher noise/reverberation were found for the younger children.

Fig. 9

Effect of noise or reverberation exposure (negative, no effect, positive) on children’s mathematical performance (including children with special educational needs (SEN)) when the task was presented in different modalities collated from reviewed papers with references in square brackets

5 The Effect of Classroom Acoustic Conditions on Primary School Children’s Listening Comprehension

This second scoping review explores further the effects of classroom acoustic conditions on auditory stimuli more generally. There have been two recent systematic reviews/meta-analyses on the effects of noise on children’s listening comprehension [10, 39]. Overall, these reviews showed poorer listening comprehension in noisy conditions. However, these reviews have included speech intelligibility/perception and memory tests in their analyses, only included classroom chatter as the noise source, and only included children who were typically developing. The aim of this scoping review was to only focus on listening comprehension as defined by Kiessling et al. as “the reception of information, meaning, or intent” [9] (p. 93) (rather than also including speech intelligibility/perception and memory), but to broaden it to any classroom acoustic conditions (i.e. internal noise, external noise, reverberation), provide a more detailed, specific analysis of these acoustic conditions, and to scope the literature of children who are not only typically developing, but also those with special educational needs.

The following research question was formulated: What is known from the literature about the effect of different classroom acoustic conditions on primary school children’s listening comprehension?

6 Method

The same PRISMA-ScR method was used for this second scoping review as the first scoping review. The database search was conducted on the 26 May 2022. The search term used was classroom AND (acoustic* OR noise OR reverb*) AND ("listening comprehension" OR "auditory comprehension" OR "spoken language comprehension" OR "speech comprehension”).

7 Results

7.1 Selection of Sources of Evidence

The search and selection process of the studies included in the review is shown in Fig. 10. After duplicates were removed, a total of 43 references were identified from searches of electronic databases. Based on the title and/or the abstract and full text, 33 papers were excluded for the following reasons: 26 did not assess the effect of classroom acoustics on children’s listening comprehension, and seven did not assess children in the primary school age range.

Fig. 10

Search strategy and results for the listening comprehension scoping review

7.2 Characteristics and Results of Sources of Evidence

A summary of the characteristics and overall outcomes of the studies included in the review is shown in Table 2. The range of publication years was from 2010 to 2021, with the majority of studies published since 2018 (see Fig. 11). A synthesis of the results follows.

Table 2 General information for the 10 papers included in the review and effect on children’s listening comprehension

Fig. 11

Publication years of the 10 studies included in the listening comprehension scoping review

7.3 Populations

The majority of studies were conducted with children from typically developing populations who had normal hearing [40,41,42,43,44,45,46,47]. One study, however, compared children who had normal hearing with those who had permanent unilateral hearing loss [48]. Another study compared children who were native Swedish listeners with non-native Swedish listeners [49]. No other special populations were studied. The full range of primary school ages were studied (i.e. 5–12 years) plus an additional study that also included 13-year-olds.

7.4 Acoustic Exposure

All studies assessed the effect of acute noise exposure during the experiment. No studies assessed the effect of chronic noise exposure. The only acoustic variable assessed was noise (used to calculate the SNR)—no studies assessed reverberation or other room acoustic variables. Different types of noises were assessed, however. These included one-competing speaker [40, 47], two-talker babble [43, 48], four-talker babble [41, 43, 47, 49], classroom noise without speech [45, 46], unintelligible classroom noise [42], traffic noise [42], and speech-shaped noise [44]. A range of presentation methods were used including headphones, loudspeakers played to individual children, and loudspeakers played to the whole class (see Table 1).

7.5 Measures and Methods

A range of different tests were used to assess children’s listening comprehension across the reviewed studies as shown in Table 2. These included children listening to a story/passage, sentences, or instructions. The types of responses ranged from answering content to inference questions, or picture-verification/matching tasks. A range of listening comprehension processes were examined; for example, in The Listening Comprehension Test 2 [50] used by Sullivan et al. [46], the child answered questions in the following domains: main idea, details, reasoning, vocabulary, and understanding messages.

7.6 Outcomes

A summary of the overall outcomes of the studies is shown in Table 2. These findings are described further in Appendix B categorised into the type of noise investigated.

Figure 12 shows a visual representation of these results categorised by the noise type. Note that the study by Griffin et al. [48] was not included as it did not compare children’s results across SNRs, but rather populations.

Fig. 12

Visual representation of the results of the reviewed studies. Lines represent when a range of SNRs were assessed. For studies that included a quiet condition without specifying the SNR, and SNR of + 25 dB was allocated for the reference condition

Overall, the majority of studies found a negative effect of SNRs below + 10 dB on children’s listening comprehension compared to quiet conditions. This was found across most of the different types of noise sources (one-competing speaker, four-talker babble, classroom noise without speech, unintelligible classroom noise, traffic noise, speech-shaped noise, but not two-talker babble). Of those studies that assessed multiple noise sources, speech distractors had a greater negative affect than classroom noise without speech [40] and unintelligible classroom noise had a greater negative affect than road traffic noise [42]. For the studies that assessed the effects for different age groups, analyses revealed a more adverse effect of noise for younger children [40, 42]. Of the studies that assessed special populations, children with unilateral hearing loss or non-native listeners were more affected than their peers from typical populations [48, 49].

For the studies that did not find an effect, there are a few nuances to note. Schiller et al. [45] did not find an effect of SNR on children’s listening comprehension; however, the authors did only assess an SNR range of + 2 to + 9 dB and did not include a quiet condition. Rudner et al. [43] used two-talker babble as the noise source at a + 10 dB SNR compared to quiet which may not have been an SNR low enough to see the effects for this type of listening comprehension task given that SNRs in classrooms range from -16 to + 23 dB [11]. This SNR was also used by Nirme et al. [41] who did not find an effect compared to quiet for inference questions, but did find a negative effect for content questions. It is possible that children could infer the answer for these questions at this SNR without having to properly hear all of the passage accurately like they needed to for the content questions. This effect of the type of question was also seen by Prodi et al. [42], who found a negative effect at − 5 dB SNR compared to quiet for the details, reasoning, vocabulary, and understanding messages domains, but no effect for the main idea domain. This suggests that higher-order comprehension tasks may be more affected by noise. The population studied is another factor to consider—Brannstrom et al. [49] did not find an effect of a 0 dB SNR compared to + 10 dB SNR on children’s listening comprehension for native listeners, but did find a negative effect for non-native listeners. A quiet condition was not included in this study, however. Additionally, the type of voice used for the stimuli can affect the results. Schiller et al. [44] did not find an effect of a 0 dB SNR compared to quiet on children listening comprehension using a normal voice for the stimuli, but did find a negative effect when a dysphonic voice was used which may reflect a teacher’s voice if they are experiencing vocal fatigue. von Lochow [47], however, did not find an effect of typical vs. dysphonic speaker in quiet or at + 5 dB SNR.

8 Discussion

The aim of the two scoping reviews was to investigate the effect of classroom acoustic conditions on children’s numeracy performance and listening comprehension. Regarding the numeracy studies, mixed results were found overall, so it is difficult to draw firm conclusions on the effect of different acoustic conditions on children’s numeracy performance. However, given that several studies did report negative effects of noise on children’s mathematical performance, it is recommended to keep noise levels in classrooms as low as possible. Shield and Dockrell [32] showed a trend of poorer mathematics performance from 30 dB upwards with it steepening from 60 dB and above so keeping noise levels below this may be ideal.

Regarding the listening comprehension studies, SNRs below + 10 dB mostly had a negative effect on children’s listening comprehension compared to quiet conditions; however, variables such as the noise type, SNR tested, the listening comprehension domain examined, the population studied, and the voice used for the stimuli affected this.

Given these findings, it would be beneficial to keep noise levels as low as possible, especially for tasks that involve listening comprehension, but this may also be beneficial for numeracy tasks. This includes minimising both internal classroom noise such as background speech from the children and movement of objects in the classroom as well as noise from outside the classroom such as road traffic noise. This can be achieved by classroom control of the children, keeping windows closed when there is risk of external noise, and installing acoustic treatment and insulation.

8.1 Limitations of the Reviewed Papers

Taking into account the findings of all of the studies included in the review, there were several limitations that were common to both the reviewed numeracy and listening comprehension papers. These create gaps in the current knowledge of how classroom acoustic conditions affect children’s numeracy performance and listening comprehension. The first limitation is the noise type used. In the numeracy papers, only a handful of studies used classroom noise as the noise source. In the listening comprehension papers, only acute noise exposure was investigated. No studies investigated the long-term effects of chronic noise exposure. Additionally, not all of the noise sources heard in the classroom have been investigated, and those that have been investigated have been at limited SNRs. Furthermore, the ecological validity of the created stimuli was often questionable. For example, many studies did not specify whether the speech stimuli and the noise were spatialised like they are in the real world which affects speech intelligibility [51]. The second limitation is the acoustic parameters investigated. In the numeracy papers, only one study investigated the effect of reverberation. In the listening comprehension papers, SNR was the only acoustic parameter investigated. No studies assessed other room acoustic parameters such as reverberation time. The third limitation is the population studied. The majority of studies focused on typically developing children. In the numeracy papers, only one study specifically assessed children with special educational needs (in this case hyperactive children) [23]. In the listening comprehension papers, only two studies explored children with special educational needs (hearing loss and non-native language listeners) [48, 49]. Additionally, none of the numeracy studies assessed children across the full primary school age range. The fourth limitation is the types of assessments used. In the numeracy papers, only two studies assessed mathematical problems presented in the auditory domain and many did not specify what domain (visual or auditory) that the stimuli were presented in. In the listening comprehension papers, a large range of tests and different listening comprehension domains were assessed using different methods making it difficult to draw conclusions across studies. Future research needs to address these limitations are outlined in the following section.

8.2 Future Research Needs

8.2.1 Noise Type

The effect of different noise types for chronic noise exposure and acute noise exposure on children’s numeracy performance would be interesting to further explore, particularly the effect of classroom noise. Noise generated by the children is becoming more of an occurrence due to modern teaching methods and modern classrooms [14,15,16,17]. Therefore, assessing children’s numeracy skills in group work activities and innovative learning environments is important.

Regarding listening comprehension, an area for future research includes examining the effect of chronic noise exposure on children’s listening comprehension. All of the reviewed studies concentrated on acute noise exposure during the experiment, but it would be interesting to see if exposure to noise while trying to comprehend speech over a long period of time (say, several years) affects children’s performance on listening comprehension tests compared to control children who have not been exposed to noise.

Additionally, different types of noise could be investigated further. The majority of studies reviewed assessed listening comprehension against one, two, or four competing speakers, or classroom noise. Only one study examined the effect of external noise and this was road traffic noise [42]. It may be beneficial to study other types of noise such as aircraft or rail noise which can be heard in classrooms and have been shown to have negative effects on other processes such as reading and cognition (see [2] and [5] for reviews). Though it is important to note from this review that speech distractors had a greater negative affect than classroom noise without speech [40] and unintelligible classroom noise had a greater negative affect than road traffic noise [42].

It would also be beneficial for future research to include ecologically valid representations of the classroom environment including spatialised stimuli like is found in the real world as this which affects speech intelligibility differently to non-spatialised stimuli [51].

8.2.2 Acoustic Parameters

Reverberation in the classroom is an acoustic feature that needs more exploration on how it affects children’s numeracy performance. Only one study in the numeracy papers review assessed reverberation time and did so using a small difference of 0.57–0.69 s. However, classroom reverberation times have been found to vary between 0.2 and 1.9 s [11]. Therefore, it would be beneficial to assess children’s numeracy performance in a larger range of reverberation times and in combination with noise. Assessing other room acoustic parameters such as early decay time (EDT), speech clarity (C50), and definition (D50) and how these may affect children’s numeracy performance may also be helpful as a significant positive correlation between reading speed and speech clarity has been found [52].

Regarding listening comprehension, it would also be worthwhile assessing the effect of these other room acoustic parameters including reverberation. It is well known that noise combined with reverberation has an even more detrimental effect on speech intelligibility [8]. Therefore, it would be beneficial to assess the effect of noise and different reverberation times on children’s listening comprehension. This would provide more evidence for what reverberation time should be achieved in classrooms.

It would also be beneficial to investigate children’s listening comprehension across a range of SNRs in the one study using the same methods. The majority of studies reviewed assessed children’s listening comprehension in one chosen SNR compared to quiet, and these chosen SNRs varied from + 10 dB to − 5 dB. Overall, it appears that children’s performance is poorer at these SNRs compared to quiet; however, it would be beneficial to assess at what SNR listening comprehension begins to be compromised and how much it is affected by lowering the SNR. For adequate speech perception, it is generally accepted that the SNR needs to be at least + 15 dB [8, 11]. It would be interesting to explore whether this is similar for listening comprehension which has the benefit of contextual cues for filling in missing information, but has higher processing demands.

8.2.3 Population

Conducting research on the effect of different acoustic conditions on children’s numeracy performance across the age range is important to assess to see if the maturation trends of older children being less affected by noise on their numeracy performance are true. Research shows these maturation effects for speech perception in noise with older children and adults having gradually better performance than younger children [53] so it would be interesting to assess if there is a maturation effect for numeracy performance in noise in the auditory domain, but also in the visual domain. Although a clear age trend was not obvious when aggregating the reviewed studies together, three studies that did examine different age groups in the same conditions in this review found a negative effect of noise on the younger children in their sample compared to the older children [31, 36, 37]. Alternatively, however, a longer duration of exposure to noise could result in an increased negative build-up effect of noise on children’s numeracy development. Therefore, further exploring the effect of noise over the schooling age span is important.

Furthermore, the socioeconomic background of the participants is important to consider as there is a medium to strong relationship between socioeconomic background and academic achievement [54]. It is important that the socioeconomic background of the participants is controlled for in these studies, especially when considering chronic noise exposure in different schools, and when conducting experimental tests with different groups of children completing the different conditions. Cohen et al. [26] raised the point that the results of their study should be taken with caution as school and district teaching policy, teaching quality, level of federal aid to a school, and school administration were not controlled in choosing the schools which may have a greater effect than noise on children’s mathematics performance. While Haines et al. [27] found a decrease in children’s mathematics performance as the noise level increased in an unadjusted analysis, once the data were adjusted for socioeconomic status (children with free school meals), the association was no longer significant. Therefore, it is very important that socioeconomic background is taken into consideration when analysing results.

Only one of the numeracy studies specifically assessed children with special educational needs (in this case hyperactive children) and found that these children performed better in low noise [23]. Johansson [24] did, however, analyse the results in terms of children with high intelligence and low intelligence and found children with low intelligence showed a trend of poorer performance in noise though this was not significant. Therefore, more research with children who have special educational needs is needed as these children may be more affected by noise than their typically developing peers. This research should include children with autism who have been found to show more repetitive behaviours such as repetitive motor movements, repetitive speech, ear covering, hitting, loud vocalisations, blinking, and verbally complaining in higher classroom noise conditions [55]. Children with English as a second language is another group that would benefit from research as they can have difficulty listening in the presence of noise and reverberation [56].

More research is also needed investigating the effect of different classroom acoustic conditions on listening comprehension for children with special educational needs. Out of the 10 reviewed studies, only two papers investigated children with special educational needs (Brannstrom et al. [49] who included non-native listeners and Griffin et al. [48] who included children with unilateral hearing loss) and both of these papers found poorer performance by the special population at adverse SNRs compared to their typical peers. Therefore, it would be beneficial to look at other populations such as children with different degrees of bilateral hearing, children with attention deficit hyperactivity disorder, and children with autism spectrum disorder to see how they perform in different acoustic conditions and compared to a control group.

8.2.4 Assessment Type

Presentation type is another factor to take into consideration when considering the effect of different classroom acoustic conditions on children’s numeracy performance. A range of test presentations were found in this review including written presentation, auditory presentation, numerical equations, and problems in word form. Exploring if there is a different effect of noise on these presentations would be interesting, for example, if there is a difference for problems expressed in numbers or in words. Additionally, the effect on problems that are presented visually in written form versus aurally would be beneficial to explore. If the presentation is in the auditory domain, then it is important to note that speech intelligibility and processing speed can be negatively affected [57, 58]. If the presentation is in the visual domain, noise can have a negative effect on children’s visual short-term memory, reading, and writing performance (see [4, 5], and [2] for reviews). In terms of outcome measures to assess, including both accuracy of solving the mathematical problem and speed of response is useful as included in several of the studies reviewed [33, 36,37,38]. Response times can be a useful measure of cognitive processing capacity [59]. Task difficulty is another factor to consider as Caviola et al. [37] found better performance in quiet and traffic noise compared to classroom noise for low difficulty problems, but no difference for high difficulty problems. A study on adults with different workplace noises showed that simple tasks are more affected by the level of the noise present, whereas difficult tasks are affected by the type of noise—in particular, intermittent and fluctuating noise distract attention and memory, more than a higher level of steady noise [60]. Although this study did not involve numeracy tasks, attention and memory are important processes for completing mathematical problems. Exploring the effects of different noise types in children for numeracy tasks varying in difficulty would therefore be of interest.

Regarding listening comprehension, it would be beneficial to assess a range of listening comprehension processes. A test that may be helpful for this is the Listening Comprehension Test 2 [50] as used by Sullivan et al. [46]. This test covers a range of listening process needed in classroom: “The test, as closely as possible, models the type of listening required in the classroom. The student must determine what part of the message needs immediate attention, organise and understand the input, and plan appropriate responses. In order to respond, the student must integrate the communication skills of vocabulary and semantics, syntax and morphology, phonology, and thinking.” [50]. This test assesses the main idea, details, reasoning, vocabulary, and understanding messages. It would therefore be beneficial to use this test across a range of SNRs using the same methods to see how the different listening comprehension domains are affected. As seen by Sullivan et al. [46], a study like this may show that higher-order comprehension tasks are more affected by noise and be able to determine at what SNRs these effects take place. This would provide more evidence for what SNRs should be achieved in classrooms.

There are a couple of methods that could be suitable when considering how to conduct future research on the effect of different classroom acoustic conditions on children’s numeracy performance and listening comprehension. Testing can be carried out directly in the real classroom which provides ecological validity. However, if the researcher is after more control in manipulating the classroom acoustic conditions, the testing could be completed in the laboratory using an auralised classroom environment. Using computer software, different classroom acoustic conditions can be generated and manipulated and then played to children while assessing their performance on numeracy tasks. This means that different acoustic conditions could be examined as well as the effect of acoustic treatment without the expense of setting these conditions up in a real classroom. Schiller et al. [45] discussed the limitations of laboratory and field methods in their paper and attempted to bridge the gap between the two. This was achieved by testing children in their habitual classrooms with their peers during school hours using stimuli presented in a diffuse field via loudspeakers rather than by headphones without spatialisation. Furthermore, Doggett et al. [61] combined an auralised classroom environment with virtual reality to assess whether acoustic treatment reduced the effects of ambient noise on cognitive performance, physiological stress, and mood in university students. The virtual headset provided a 360 degree view of a classroom, while conditions consisting of no noise, untreated room noise, and treated room noise were played. The authors concluded that this virtual reality set-up was an effective and efficient way to evaluate the effects of different acoustic conditions on university students’ cognition. While this study was conducted with university students, advancements in technology can potentially provide creative, realistic options of assessing the impact of different classroom acoustic conditions on children’s classroom performance. There is still room for making the task and stimuli more representative of the real classroom environment. Building on this method of bridging the gap between the laboratory and the field in the future will be beneficial for making sure assessments are conducted in an ecologically valid way that transfers to the real classroom acoustic environment so recommendations on the acoustic conditions needed for accurate numeracy performance and listening comprehension can be made.

9 Conclusions

The first scoping review synthesised research that assessed the effect of different classroom acoustic conditions on primary school children’s numeracy performance. Overall, the majority of studies showed a negative effect or no effect of noise on mathematics performance, however, there were a couple of studies that showed a positive effect. This motivates the need for future research assessing the effects of different noise types and levels on children in different years of primary school to gather a clearer picture of the effect. As well as assessing the effect of age and noise type, this future research could also include controlled studies of reverberation, numeracy problem presentation type, and the impact on children with different special educational needs. Although definitive conclusions could not be drawn on the impact of noise on children’s numeracy performance, the fact that many studies did show a negative effect demonstrates that it would be beneficial to keep classroom noise levels low as a precaution.

The second scoping review provided a synthesis of what is known from the literature about the effect of different classroom acoustic conditions on primary school children’s listening comprehension. The majority of reviewed studies used babble or classroom noise as the noise source. Overall, it was found that SNRs below + 10 dB mostly have a negative effect on children’s listening comprehension compared to quiet conditions, however, variables such as the noise type, SNR tested, the listening comprehension domain examined, the population studied, and the voice used for the stimuli, can affect this. This review identified important gaps in knowledge of how the acoustic conditions of classrooms affect children’s listening comprehension and proposed future research ideas to help fill these gaps.

Overall, these two scoping reviews showed that poor classroom acoustic conditions can have a negative effect on children’s listening comprehension in particular, but can also affect their numeracy performance. However, many future research opportunities were revealed due to some mixed results and gaps in the current literature.

Appendices

Appendix A: Summary of Outcomes for Numeracy Studies

1.1 Chronic Noise Exposure

Two studies assessed the effect of aircraft noise on children’s mathematics achievement. Cohen et al. [26] found that children in noise-abated schools had significantly better scores on the mathematics achievement test compared to both the schools with high aircraft noise and low aircraft noise. However, these results should be taken with caution as the authors note that school and district teaching policy, teaching quality, level of federal aid to a school, and school administration were not controlled in choosing the schools which may have a greater effect than noise on children’s mathematics performance. Haines et al. [27] found that children’s mathematics performance decreased by 0.73 of a mark as the noise level increased in each 3 dB contour band, However, once the data were adjusted for socioeconomic status (based on children with free school meals), the association was no longer significant. These results show that aircraft noise may negatively affect children’s mathematical performance; however, it is important when designing such chronic noise exposure studies that the socioeconomic status and other educational factors are matched as closely as possible between noise-exposed schools and control schools as these factors may influence the results.

Two studies assessed the effect of road traffic noise on children’s mathematics achievement. Papanikolaou et al. [28] found that children from schools with low traffic noise levels had significantly better scores than children in high-level noise schools. Additionally, children from low-level noise schools and medium-level noise schools had significantly higher task completion rate than children in high-level noise schools. All schools had a similar socioeconomic background. These results show the negative effect that chronic exposure to everyday road traffic noise at levels of 72–80 dB can have on children's mathematical performance. Shield and Dockrell [32] also assessed the effect of external noise (mainly road traffic noise, range = 30–80 dB L_Aeq) and found a significant negative relationship between external noise and Year 2 and Year 6 children’s mathematics scores. These results are consistent with Papanikolaou et al. [28] demonstrating the negative effect of chronic exposure to road traffic noise on children’s mathematical performance.

Two studies assessed the effect of heating and/or cooling mechanical noise on children’s mathematics achievement. No significant correlations between mathematics test score and noise level were found in either the study by Ronsse and Wang [29] or Ronsse and Wang [30]. These results suggest that chronic exposure to the ambient noise from heating and/or cooling systems in the range of 34–55 dBA does not affect children’s mathematical performance.

Two studies assessed the effect of occupied classroom noise on children’s mathematics achievement. Shield and Dockrell [31] found that occupied L_A90 noise levels were negatively correlated with Year 2 children’s mathematics scores, but not Year 6 children’s mathematics scores. Shield and Dockrell [32] found a negative relationship between internal noise level and children’s mathematics scores; however, this was not significant. These results are mixed but suggest that occupied classroom noise may affect mathematical performance in younger children.

1.2 Acute Noise Exposure

Six studies included babble, intelligible classroom noise, or irrelevant classroom noise as the noise source of acute exposure in an experimental format. Negative effects of noise were found in four studies (though some results depended on the population characteristics). Dockrell and Shield [34] found that children performed significantly better in a quiet condition than a children’s babble condition. Performance in a babble and environmental noise condition was not statistically significantly different to that in the quiet condition or babble condition. Different groups of children completed the different noise conditions; however, so group differences may have contributed to these results. Meinhardt-Injac et al. [36] found younger children showed poorer accuracy during irrelevant speech (foreign language reading of a Danish newspaper article) compared to the baseline condition. However, for older children, there was not a significant effect of irrelevant speech or irrelevant classroom noise (typical classroom sounds without speech) on the proportion of correct responses compared to a pink noise baseline condition. The same result was found when analysing the children’s reaction time. Zentall and Shaw [23] found that hyperactive children performed poorer in a high noise condition of intelligible classroom noise (real recordings of grade two children during free time with high linguistic content from the children and teacher speaking) compared to a low noise condition; however, control children that were not hyperactive performed significantly better in the high noise condition than in low noise condition. There were no significant differences between the two groups in terms of age, IQ, or achievement. Caviola et al. [37] found that 11-year-olds had significantly better aural mental calculation in quiet and traffic noise compared to irrelevant classroom noise (typical working classroom sounds mixed with a fluctuating standard noise signal of Italian speech which is unintelligible). For the 12-year-olds, there were significantly better scores in quiet only compared to irrelevant classroom noise. For the 13-year-olds, no statistically significant difference between listening conditions was found. In terms of the difficulty of the mathematical equation, the authors found better performance in quiet and traffic noise compared to irrelevant classroom noise for low difficulty problems, but no difference for high difficulty problems. Regarding response times, the authors did not report the effect of listening condition by age group. They did report the interaction between listening condition and level of difficulty, with the results the same as for accuracy, i.e. faster performance in quiet and traffic noise compared to classroom noise for low difficulty problems, but no difference for high difficulty problems. These results show that babble, intelligible classroom noise, and irrelevant classroom noise can have a detrimental effect on children’s mathematical performance, especially for younger children or hyperactive children. The finding of Zentall and Shaw [23] that control children performed significantly better in the high noise condition than in low noise condition, however, was interesting and is unpacked more in the discussion.

Two studies did not find effects of irrelevant speech or intelligible speech noise on mathematical performance. Ljung et al. [35] found no significant differences in performance between quiet, road traffic noise or irrelevant speech (unintelligible babble and intelligible conversation segments) conditions. Kassinove [33] found that there was no effect of any of the noise conditions [i.e. (1) quiet; (2) stories; (3) popular music; (4) music and stories simultaneously presented from the same source; or (5) music and stories simultaneously presented from the different sources] on the children’s mean time per response, the variability of response times, the number of correct responses, or the probability of error (number wrong divided by number attempted). The change in response time over the 45 min assessment was not significant either. The only significant finding was that children tended to have more time-outs (i.e. a period where the child turned away from the task and looked to tune out for 2 s or more) the longer the task went for. These results conflict with the results in the previous paragraph that show a negative effect of intelligible or irrelevant speech noise on children’s mathematical performance. Therefore, the overall finding for this speech/classroom noise source category is mixed and inconclusive.

One study assessed children’s mathematical performance in octave band noise. Johansson [24] found that children with high intelligence solved more items in continuous and intermittent noise than in quiet conditions. However, this was not the case for children with low intelligence who showed a trend of poorer performance in noise though this was not significant. Different groups of children completed the different noise conditions, however, so group differences may have contributed to these results. These results show that the effect of octave band noise on mathematical performance may be different for different children.

1.3 Acute Reverberation Exposure

Finally, one study assessed the effect of reverberation of children’s mental calculation of an aurally presented stimulus. Prodi and Visentin [38] found that the effect of reverberation on accuracy was only apparent in the classroom noise condition and depended on the grade of the children with younger children performing poorer than older children in the longer reverberation condition but not the shorter condition. No significant effects of reverberation and listening condition on reaction time were found. Different groups of children completed the different reverberation conditions, however, so group differences may have contributed to these results. This result shows the negative impact that noise combined with a longer reverberation time can have on children’s mathematical performance, particularly for younger children.

Appendix B: Summary of Outcomes for Listening Comprehension Studies

1.1 Two-Talker Babble

Rudner et al. [43] conducted four studies investigating background noise, voice quality, and visual cues effects on children’s listening comprehension. For Study 4, the four conditions investigated were: (1) audio-only in quiet; (2) audio-only in two-talker multi-babble child noise (native, meaningful) (+ 10 dB SNR); (3) two-talker multi-babble child noise (native, meaningful) (+ 10 dB SNR) with congruent visual support; and (4) two-talker multi-babble child noise (native, meaningful) (+ 10 dB SNR) with visual information that was incongruent with the noise. The authors, however, did not find any significant differences between conditions.

Griffin et al. [48] assessed listening comprehension in children with normal hearing and children with unilateral hearing loss in quiet, two-talker babble (native but non-meaningful) with a + 6 dB SNR, and two-talker babble with an individualised SNR required to achieve 50% sentence understanding. The purpose of this paper was to compare performance across populations rather than SNRs so this analysis was not conducted. (And there would have been issues with the presentations not being counterbalanced.) The analyses that were conducted showed that children with unilateral hearing loss perform similarly to children with normal hearing in favourable listening conditions (i.e. quiet, + 6 dB SNR), but in challenging listening conditions (individualised SNR), many children with unilateral hearing loss performed more poorly than the children with normal hearing.

1.2 Four-Talker Babble

Brannstrom et al. [49] assessed the effect of four-talker child babble noise (native and meaningful) on children’s listening comprehension at a typical SNR (0 dB SNR) and a favourable SNR (+ 10 dB SNR). A negative effect of the lower SNR was found for non-native listeners, but there was no effect for native listeners.

Nirme et al. [41] assessed children’s listening comprehension in audio-only in quiet, audio-only in four-child multi-talker babble noise (native, meaningful) (+ 10 dB SNR), audio-visual in quiet, and audio-visual in four-child multi-talker babble noise (native, meaningful) (+ 10 dB SNR). The authors found a negative effect of noise for content question on the listening comprehension test, but no effect for the inference question. Visual presentation of the stimuli by the virtual speaker showed a marginally significant effect of reducing the effect of noise compared to audio-only. The child’s executive functioning helped their listening comprehension in quiet, but not in noise.

Rudner et al. [43] also investigated the effect of four listening conditions on children’s listening comprehension: (1) audio-only in quiet, (2) audio-only in four-talker multi-babble child noise (native, meaningful) (+ 10 dB SNR), (3) audio-visual iin quiet, and (4) audio-only in four-talker multi-babble child noise (native, meaningful) (+ 10 dB SNR). There was a significant negative effect of four-talker multi-babble child noise on children’s listening comprehension.

von Lochow [47] assessed children’s listening comprehension in six conditions: (1) typical voice in quiet; (2) typical voice with one competing child speaker (native, meaningful) (+ 5 dB SNR); (3) typical voice with four competing child speakers (native, meaningful) (+ 5 dB SNR); (4) dysphonic voice in quiet; (5) dysphonic voice with one competing child speaker (native, meaningful) (+ 5 dB SNR); and (6) dysphonic voice with four competing child speakers (native, meaningful) (+ 5 dB SNR). The authors found a significant effect of background noise but no effect of the number of competing speakers. There was no effect of typical vs. dysphonic speaker.

1.3 Classroom Noise

Klatte et al. [40] assessed children’s listening comprehension in quiet, non-native one-talker background speech (− 3 to + 3 dB SNR depending on seating location), and classroom noise without speech (− 3 to + 4 dB SNR depending on seating location). Children’s listening comprehension was significantly compromised by background speech and classroom noise, and this effect was stronger for first than third grade children. Background speech had a stronger effect than classroom noise.

Schiller et al. [45] assessed children’s listening comprehension in created classroom noise (no speech) conditions between + 2 and + 9 dB SNR. The authors found no difference in the children’s listening comprehension across SNRs.

Sullivan et al. [46] assessed children’s listening comprehension in quiet and multi-classroom noise (− 5 dB SNR) (no more information on what the multi-classroom noise consisted of was provided). Overall, noise had a significant negative effect on children’s listening comprehension. When explored further into the types of domains affected, significant negative effects of noise were found for the details, reasoning, vocabulary, and understanding messages domains, but no effect of noise was found for the main idea domain. The authors suggest that a lower vocabulary size and working memory capacity may contribute to poor listening comprehension in noise.

Prodi et al. [42] assessed children’s listening comprehension in quiet, road traffic noise (0 dB SNR), and unintelligible classroom noise (0 dB SNR). The authors found that sentence comprehension accuracy was negatively affected by noise when ages were combined. Accuracy, however, was close to ceiling for all listening conditions. Analyses of the children’s reaction times showed longer processing times for classroom noise, then traffic, compared to quiet for the 10-year-olds and 11-year-olds. For the 12- and 13-year-olds, the same effect as for the younger children was seen for the more complex sentences, but there was no effect of noise for the simpler sentences.

1.4 Speech-Shaped Noise

Schiller et al. [44] assessed children’s listening comprehension using a normal voice in quiet (+ 31 to + 33 dB SNR), an impaired voice in quiet (+ 31 to + 33 dB SNR), a normal voice in speech-shaped noise (0 dB SNR), and an impaired voice in speech-shaped noise (0 dB SNR). The authors found a negative effect of noise when the stimuli were presented in a dysphonic voice; however, there was no effect of noise when the stimuli were presented in a normal voice.