Noise Health Home 

ARTICLE
[Download PDF]
Year : 2013  |  Volume : 15  |  Issue : 62  |  Page : 73--78

Mental arithmetic and non-speech office noise: An exploration of interference-by-content

Nick Perham1, Helen Hodgetts2, Simon Banbury3,  
1 Department of Applied Psychology, Cardiff Metropolitan University, Cardiff, United Kingdom
2 School of Psychology, Cardiff University, Cardiff, United Kingdom
3 CAE, Professional Services, Ottawa, Canada

Correspondence Address:
Nick Perham
Department of Applied Psychology, Cardiff Metropolitan University, Cardiff CF5 2YB
United Kingdom

Abstract

An interference-by-content account of auditory distraction - in which the impairment to task performance derives from the similarity of what is being recalled and what is being ignored - was explored concerning mental arithmetic performance. Participants completed both a serial recall and a mental arithmetic task in the presence of quiet, office noise with speech (OS) and office noise without speech (ONS). Both tasks revealed that the two office noise condition's significantly impaired performance. That the ONS produced this deficit suggests that an interference-by-content account cannot explain impairment to mental arithmetic performance by background sound.



How to cite this article:
Perham N, Hodgetts H, Banbury S. Mental arithmetic and non-speech office noise: An exploration of interference-by-content.Noise Health 2013;15:73-78


How to cite this URL:
Perham N, Hodgetts H, Banbury S. Mental arithmetic and non-speech office noise: An exploration of interference-by-content. Noise Health [serial online] 2013 [cited 2019 May 22 ];15:73-78
Available from: http://www.noiseandhealth.org/text.asp?2013/15/62/73/107160


Full Text

 Introduction



Performing tasks in the presence of background sound is an almost daily occurrence for many people and it can have negative effects on both health and performance. [1],[2] One of the most robust and explored laboratory paradigms that has sought to emulate this phenomenon is the irrelevant sound effect (ISE). This simple task shows that serial recall performance is poorer in a background sound compared to a quiet control condition. Two broad classes of explanations have been identified. [3] The first proposes that the decrease in performance is a result of the similarity between the to-be-ignored (TBI) and to-be-recalled (TBR) information whereas the second explanation suggests that it is a conflict with the way that the TBI and TBR information is being processed. Subsequent studies provide strong support for this latter explanation. [4] Further studies reveal that other tasks are also susceptible to this phenomenon - free recall, train timetable, language learning and mental arithmetic. [5],[6],[7],[8] Yet very little further work has been conducted upon them with regard to the possible content versus process explanations. This study explores whether the content explanation of interference can explain mental arithmetic in the presence of background sound by comparing mental arithmetic performance in the presence of non-speech office noise against performance in quiet. According to the content approaches, the non-speech office noise should not disrupt performance due to the lack of similar phonological sounds, whereas, the process approach argues that it should due to there being a conflict of process. Given that this non-speech office sound is quite novel, serial recall performance under this sound condition is first explored to indicate that the sound produces an ISE. The second study tests whether the sound impairs mental arithmetic performance.

The ISE typically involves serial recall in which visually-presented items (around 7-9 digits or consonants) are presented one at a time and then recalled in their strict presentation order whilst participants are instructed to ignore any background sound that is usually presented through headphones. For the ISE to be observed, two characteristics must be present. The first is that the background sound must contain acoustical variation (termed changing-state). That is, the sound must show appreciable acoustical difference between each successive sound item. For example, disruption is far greater with an irrelevant auditory sequence such as "n, r, p" but not when the sequence is 'steady-state' as in "c, c, c". The second characteristic is that the task itself must require the use of seriation - that is, a need for the participant to use rehearsal as a means of retaining and retrieving the order of presented information. [9] When this requirement is not present, as in the missing item task or when items are recalled by categories, then the ISE is absent. [10],[11] Furthermore, the ISE cannot be habituated e.g., [12],[13] is not affected by preference e.g., [14],[15] affects around 92% of participants. [16] and it is independent of the intensity of the sound (from the levels of a whisper, 48dB[A], up to a shout, 76dB[A], [17] ).

Explanations of the ISE have been broadly categorized as either interference-by-content or interference-by-process accounts. [3] The former encapsulate a structuralist approach in assuming a limited capacity for short-term memory performance. Examples of the former are the working memory model, [18] the feature model [19] and the attentional capture account. [20] The working memory model stems from the hugely influential account of human memory proposed by. [21] From subsequent revisions, the model explains the ISE as resulting from confusion between phonological information in the TBR and TBI. However, many studies reveal that the ISE is observed when there is no such information in the TBI. [22] The feature model once again argues that disruption arises from the identity of the TBR and TBI items. However, this time the disruption is due to mismatching information, or features in the authors' terminology. If the TBI item contains items that mismatch the features of the TBR item, then the former features are 'adopted' and thus incorporated into the representation of the TBR item, resulting in poor recall. To explain the classic ISE (that of poorer performance in changing- compared to steady-state sound) recourse to an 'attention' construct is made by which changing-state sounds capture attention more than do steady-state sounds. Criticisms of this account are that, despite the authors arguing for a distinction between interference from speech and non-speech, both sounds elicit similar amounts of disruption. A second criticism can be leveled at the attentional capture account as well. The attentional capture account stems from physiological research [23] and contends that each new sound captures attention via an orienting response. [20] Thus, changing-state sounds will capture attention but steady-state sounds will not. Unfortunately, as with the feature model, it assumes that the sound will capture attention regardless of what task the participant is engaged with. However, studies show that this is not the case. For example, when recalling according to the categories that items belong to (category recall), contrary to their presentation order (serial recall), performance is not affected by changing-state sound. [11]

Mental arithmetic problems have been argued to be solved through two routes. The first route requires the participant to retrieve the answer from their long-term memory such as 'knowing' that '2 + 3 = 5'. The second route depends upon non-retrieval processes (termed procedural) such as counting and transformation. During counting, the participant reaches their answer by ascending or descending the numerical scale. [24] Transformation involves changing the original problem so that it resembles smaller, easier ones. So, in calculating '24 + 9', it is possible to transform the problem to '24 + 6 + 3' by virtue of recognizing that '6' and '3' equal '9'. After this, one can add the '6' to the '24' to make '30', and then add the remaining '3' to make '33'. The main explanation for procedural strategies invokes elements of [21] working memory model, in particular, the central executive and phonological loop. The central executive is thought to be involved with both retrieval and procedural strategies and helps plan and switch between activities. [24],[25],[26],[27] The phonological loop uses the mechanism of rehearsal to maintain verbal or phonological information and only during procedural, and not retrieval, strategies. [24] Rehearsal is used to monitor running totals when participants move around the numerical scale as well as storing results, if necessary. [24],[28]

Despite mental arithmetic performance being shown to be negatively affected by irrelevant sound in a similar way to serial recall, [1],[29],[30] little research has explored how the two interference accounts can explain performance. At first blush, it might seem that interference-by-process is an adequate explanation as [1] found that both office noise with and without speech significantly impaired mental arithmetic performance. However, there are some issues that mean that further investigation of this effect is warranted. Firstly, the mental arithmetic task used by Banbury and Berry was self-paced. This meant that participants were able to consider their total before receiving the next integer to add (or subtract) to their running total. In doing so it is highly likely that their need to use rehearsal was much reduced or even removed. Studies show that when rehearsal is not needed, such as recalling items according to categories or having all the information present whilst responding, [7],[11],[31],[32] then impairment by background sound is not observed. Thus, to fully explore whether an interference-by-process or - content account explains mental arithmetic performance better, the situation in which these explanations have been applied previously - the ISE - has to be followed. As the ISE requires the use of seriation, via the process of rehearsal, during its primary task, then it is crucial that the mental arithmetic task encourages seriation, and thus rehearsal, as well. In order to do this, we set the presentation rate of the stimuli (numbers and signs) at the rate of one every one and a half seconds. This was determined by a pilot study comparing performance at 1 s and 1½ s presentation rates - the latter elicited performance that was only around chance in the presence of no sound. The presentation rate for the serial recall task was consistent with previous studies at one item every second. [9] The second issue was the changing-state nature of the office sound. Although there is currently no means to guarantee that a sound contains enough changing-state information, one possibility is to ensure that it actually disrupts serial recall performance. As Banbury and Berry did not do this, we required that participants perform a serial recall task under the same background sound conditions as the mental arithmetic task. Thus, the current study sought to clarify and explore whether an interference-by-content account of auditory distraction could be applied to mental arithmetic performance by asking participants to performance a series of running total problems in the presence of quiet, office noise with speech (OS) and office noise without speech (ONS).

 Methods



Participants

Participants were 36 undergraduates from a south Wales University aged 18-30 years who participated in exchange for course credit or a small honorarium. All reported normal hearing and vision and were native English speakers.

Design

For the serial recall task a repeated-measures design was employed with the variables of sound (quiet, [OS] and [ONS]) and position (1-8). The mental arithmetic task only contained the variable of sound (quiet, [OS] and [ONS]). In the former task the dependent variable was the number of items recalled in their correct serial position (out of 8) and in the later task the dependent variable was the number of correct mental arithmetic problems solved (out of 10). Half of the participants completed the serial recall task and the other half completed the mental arithmetic task first.

 Materials



For the serial recall task, 20 trials of eight-item digit lists were created and divided into two blocks (one for each sound condition) with all items taken from the list one to nine and presented at the rate of one per second. No item was repeated within any single list and no four-digit sequence within a list was started by the digits '1' and '9' to prevent easier rehearsal of year information such as '1971'. For the mental arithmetic task, 20 addition running total mental arithmetic problems were created each comprising the digits 1-9 and with all final totals ranging from 60 to 99. All stimuli (numbers and signs) were presented at the rate of one every one and a half seconds. The trials for both tasks were divided into two groups of ten and fully counterbalanced with respect to the two sound conditions. These were created via Powerpoint and presented via a Samsung Syncmaster 171S PC.

Sounds were recorded using Sound Forge and mixed using the Sound Edit Pro software package within the range 65 dB (A) to 75 dB (A). The ONS sound condition contained computer humming, doors opening and closing, typing, photocopying, printing, telephone ringing, mobile telephone ringing, papers being shuffled, footsteps and knocking at the door. The OS sound condition consisted of those sounds described above as well as male and female voices reading short passages of text unrelated to the primary tasks, about such topics as knitting, ships, seeds, bread making and drums. Each speech passage lasted for between 30 s and 50 s with around half of it overlapping the above sounds and half of it not, and on average, there was one passage for every minute of sound. On some occasions, as in real life, there was silence although this only lasted a second or two. All mixes were within the range 65 dB (A) to 75 dB (A).

 Procedure



Participants were tested individually or in small groups of up to six and were seated at a viewing distance of approximately 60 cm from the PC monitor. For the serial recall task, standardized instructions informed participants that they were to be presented with twenty lists of eight items presented for 700 ms with 300 ms between each item and that their task was to recall each list in the order in which the items were presented when they viewed a 'RECALL' sign on the screen which would appear 10 s after the presentation of the final item. During some trials, the participants may hear the sound through the headphones which they were instructed to ignore. Two practice trials in the presence of silence were provided before the experiment proper.

For the mental arithmetic task, they then read standardized instructions informing them that they would receive a series of addition mental arithmetic problems in which 15 digits, and 14 addition signs, would be presented one at a time on a screen for 1 s each. Their task was to keep a running total in mind and write down the answer when shown the 'equals' sign, which appeared 10 s following presentation of the 15 th digit, whilst ignoring any sound played through headphones. Two practice trials in silence were provided before the experiment proper.

 Results



Serial recall

Strict serial recall criteria were employed so that a recalled item was only scored as correct if it appeared in the exact position in which it was presented in the list. [Figure 1] showed that in all conditions, participants demonstrated the typical serial recall curve - that is, performance was better for the early and late presented items and poorest for items presented in the middle positions. A one-tailed, two-way ANOVA showed a significant main of effect of sound, F (2, 46) = 21.26, MSE = 1.96, P < 0.001, with Least Squared Difference (LSD) pairwise comparisons showing that all sound conditions were significantly different from each other ( P < 0.01). A significant main effect of position, F (7, 161) = 19.7, MSE = 0.46, P < 0.01, was also observed. Finally, a significant sound by position was noted, F (14, 322) = 2.35, MSE = 0.02, P < 0.01. Post-hoc analyses for the main effect of position and simple effects analyses for the interaction were not conducted as [Figure 1]. clearly shows that any differences between items are explained by recourse to the serial recall curve typically observed in ISE studies and that this pattern is evident for all three sound conditions.{Figure 1}

Mental arithmetic

The pattern of data showed that performance was highest under the quiet condition (mean 6.11; standard error (SE) 0.39), then the ONS condition (mean 5.5; SE 0.44) and finally, the OS condition (mean 5.0; SE 0.04). A one-tailed, one-way ANOVA revealed a significant main effect of sound, F (2, 70) = 5.93, MSE = 0.12, P < 0.01, with post-hoc LSD comparisons showing that performance in quiet was significantly better than in ONS and OS ( P < 0.05) however there was no significant difference between performance in either of the office sound conditions.

 Discussion



The current study explored whether an interference-by-content account of auditory distraction - in which impairment derives from the similarity in content between information used in the primary task and information that is ignored in the irrelevant sound - could explain the impairment to mental arithmetic performance by non-speech (ONS) background sound. Results revealed that both office noise with and without speech (OS and ONS) impaired mental arithmetic performance. Therefore, the observation of impairment by the latter condition casts doubt on the sufficiency of the interference-by-content accounts of auditory distraction in explaining mental arithmetic performance. Further, the fact that both office sounds significantly impaired serial recall performance means we can be quite confident that both office sounds contain sufficient changing-state information to elicit an irrelevant sound effect ISE.

Although the ISE is typically demonstrated using serial recall, evidence of its effects are observed with other tasks that require seriation to successfully perform the task such as free recall, identifying the best option, mental arithmetic and language learning. [1],[6],[7],[8] Explanations of the ISE have been broadly categorized as interference-by-process and - content. [3] The main distinction between them is that the former suggests that it is the similarity between the processing of the TBR information and the TBI information that causes the impairment whereas the latter argues that it is the similarity of the content of both pieces of information that is crucial. Much evidence favours the interference-by-process account as the ISE is still observed when the TBR bears no relation, in terms of content, with the TBI [22] and also that the ISE is not observed in some situations even when the content is similar. [33]

This study attempted to explore whether an interference-by-content explanation could be applied to mental arithmetic as [1] revealed that both speech and non-speech impaired performance compared to a quiet control condition. However, as their task was self-paced, which meant that participants could take as long as they wanted before requesting the next piece of information, it was likely that participants did not need to engage in rehearsal to retain the order of information. This is problematic for delineating between the two general explanations of the ISE as it is necessary that the task requires the use of rehearsal in order to retain the order of information as the serial recall task does. To overcome this, we amended Banbury and Berry's task by setting the presentation of each item and one and a half seconds. The fact that ONS impaired mental arithmetic performance directly contradicts the working memory model explanation of the ISE [18] - an interference-by-content account. The working memory model argues that impairment derives from the confusion between phonological information in the both the TBR and the TBI. However, the ONS condition comprised a variety of office sounds, such as computer humming, doors opening and closing, typing, photocopying, printing, telephone ringing, mobile telephone ringing, papers being shuffled, footsteps and knocking at the door (the addition of narrated passages of text made the OS condition), which arguably do not contain phonological information.

Other interference-by-content accounts are no more successful in explaining the current results, especially when other recent findings are considered. The feature model claims that speech sounds are more disruptive than non-speech sounds and, at first blush, this seems consistent with the findings of the serial recall experiment in the current paper - performances in both office sounds were significantly poorer than quiet and significantly different from each other. However, in the mental arithmetic experiment, although performance in both office sounds was significantly poorer than in quiet there was no significant difference between them. Proponents of the feature model may argue that their attentional parameter, a, could be adjusted so that it predicts differences between OS and ONS on serial recall but not on mental arithmetic. This ad hoc explanation would seem to lack discernable reason why one task would elicit this pattern of impairment and another would not.

The attentional capture account also has the same difficulty in explaining why OS and ONS are significantly different for one task, serial recall, but not another, mental arithmetic. With regards to explaining the ISE, the attentional capture account proposes that certain sounds are more attention-grabbing and, as a result of this, reduce performance. [20] However, studies show that the same sounds have different effects dependent upon on the task requirements. For example, when participants are instructed to recall in list items in serial order, disruption by changing-state sound is observed yet when the instructions ask for category recall, no such disruption is noted. [11]

We favor the interference-by-process explanation and feel that it can address the issues raised above. On this account, the ISE occurs when the processing of information required in both the TBR and TBI items is the same. In the case of the ISE, this is the processing of order information and is processed in the TBR items using the process of seriation - the ability to retain information in order. In order to do so, participants rely on their language abilities through the technique of rehearsal. Order information in the background sound derives from our ability to pre-attentively process acoustical changes from one segmentable sound item to another. [4] These segmentable elements within the sound yield cues that inform the listener as to the order of those elements. [34],[35] For the ISE, the impairment occurs when the irrelevant order cues in the sound conflict with the deliberate seriation of items in the recall list. More recently, the interference-by-process account has been extended beyond the ISE and to semantic auditory distraction. The same underlying principles apply but this time the processing is of semantic, and not order, information. In a series of systematic studies, [33],[36] showed that when TBR items are recalled using semantic processing (by requiring the participants to recall a list of categorized words using free recall) they were more impaired by forward speech, compared to reversed speech, and items related, rather than unrelated, to the TBR items.

With regard to the issue of the difference between ONS and OS on mental arithmetic performance, an explanation may lie with the requirements of each task. Given that the interference-by-process account for serial recall (the ISE) and mental arithmetic depends upon the tasks requiring the use of seriation, it is perhaps no surprise that serial recall shows differences between the ONS and OS sounds for two reasons. Firstly, serial recall is arguably the purest test of seriation as the task involves recalling items in order. One might argue that there is an element of item information required yet typically in serial recall studies each trial contains the same items (usually digits 1-9 or consonants). Not only are these items overly familiar to participants, but they are repeated in each trial. Thus, if a participant cannot remember what item to write in a particular position, they may be able to guess quite accurately especially if they have already completed a reasonable proportion of the list. In contrast, in the mental arithmetic task the correct responses are not used as frequently by participants as those used in the serial recall task (e.g., "24", "33" or "85" compared with "3", "5" or "8") nor are they repeated throughout the trials. If a participant did guess the answer, then they would be very unlikely to be correct. Further, participants may use additional processes, such as problem encoding, accessing and searching long-term memory [24] which are likely to be used during mental arithmetic and these may not be susceptible to impairment from changing-state background sound. The second reason is that speech sounds generally contain more changing-state information than non-speech sounds due to the changes in acoustical variation [37] showed that speech disrupted serial recall performance more than tones. Speech has greater onset and offset times than non-speech sounds, which produce greater cues to segmentation and this results in more changing-state information. Combining both less reliance on seriation in the mental arithmetic task and the greater changing-state information in speech sounds explains why ONS and OS did not differ on the mental arithmetic task.

Recent research demonstrates an often experienced but little researched phenomenon whereby participants were more disrupted during mental arithmetic calculations when the irrelevant sound comprised numbers that were numerically similar, compared to dissimilar, to the correct answer to the problem (e.g., if the correct response was '37' the similar irrelevant numbers would be '36, 37, 38, 39, 40'). [7] At first blush this finding can be explained by both the interference-by-content and - process accounts. To delineate between these two accounts, it may be necessary to explore a direction effect, that is, whether the same effect occurs when the direction of the irrelevant numbers is the same as the direction of the calculations in the mental arithmetic task. If it only occurs when, for example, ascending irrelevant numbers during addition mental arithmetic then this would rule out an interference-by-content account and favor a - process account.

Individual differences may also play a role in how well people are able to overcome the deleterious effects of irrelevant sound when performing mental arithmetic calculations. The degree to which one engages in rehearsal to retain order information may determine how much interference is suffered when changing-state background sound is played. Indeed, recent research shows that poor readers, who arguably find it difficult to retain order information, do not demonstrate the syntax effect - poorer recall performance of syntactically-congruent items (e.g., adjective-noun pairs) compared with syntactically-incongruent pairs (e.g., noun-adjective pairs). [38],[39]

In summary, the current study shows that non-speech sounds in the form of ONS, impairs mental arithmetic performance, which is difficult to explain by recourse to interference-by-content accounts of the ISE.

 Acknowledgments



The research reported in this article received financial support from the United Kingdom's Economic and Social Research Council, reference number R000239850, awarded to Dr. Simon P. Banbury and Prof. Dylan M. Jones. Correspondence can be addressed to Nick Perham at the Department of Applied Psychology, Cardiff Metropolitan University, Cardiff, CF52YB, United Kingdom.

References

1Banbury S, Berry DC. Disruption of office-related tasks by speech and office noise. Br J Psychol 1998;89:499-517.
2Pilkington E. Buildings of future returning to nature. Guard April 1 st , 1995. p. 2.
3Jones DM, Tremblay S. Interference in memory by process or content? A reply to Neath (2000) Psychon Bull Rev 2000;7:550-8.
4Jones DM, Macken WJ, Hughes RW. Perceptual organization masquerading as phonological storage: Further support for a perceptual-gestural view of short-term memory. J Mem Lang 2006;54:265-81.
5Banbury S, Berry DC. Habituation and dishabituation to speech and office noise. J Exp Psychol Appl 1997;3:181-95.
6Beaman CP, Jones DM. Irrelevant sound disrupts order information in free recall as in serial recall. Q J Exp Psychol A 1998;51:615-36.
7Perham N, Banbury S. The role of rehearsal in a novel call center-type task. Noise Health 2012;14:1-5.
8Saffran JR, Newport EL, Aslin RN. Word segmentation: The role of distributional cues. J Mem Lang 1996;35:606-21.
9Jones DM. The cognitive psychology of auditory distraction: The 1997 BPS Broadbent Lecture. Br J Psychol 1999;90:167-87.
10Beaman CP, Jones DM. Role of serial order in the irrelevant speech effect: Tests of the changing-state hypothesis. J Exp Psychol Learn Mem Cogn 1997;23:459-71.
11Perham N, Banbury SP, Jones DM. Reduction in auditory distraction by retrieval strategy. Memory 2007;15:465-73.
12Jones DM, Macken WJ, Mosdell N. The role of habituation in the disruption of recall performance by irrelevant sound. Br J Psychol 1997;88:549-64.
13Perham N, Banbury SP. You cannot ignore it: Attention to 'irrelevant' sound during a habituation period does not produce habituation. Proceedings of the 51 st Annual Meeting of the Human Factors and Ergonomics Society. New York: NY: HFES; 2008. p. 1622-6.
14Perham N, Vizard J. Can preference for background music mediate the irrelevant sound effect?. Appl Cogn Psychol 2010;25:625-31.
15Perham N, Sykora M. Disliked music can be better for performance than liked music. Appl Cogn Psychol 2012;26:550-5.
16Ellermeier W, Zimmer K. Individual differences in susceptibility to the "irrelevant speech effect". J Acoust Soc Am 1997;102:2191-9.
17Colle HA. Auditory encoding in visual short-term memory: Effects of noise intensity and spatial location. J Verbal Learning Verbal Behav 1980;19:722-35.
18Baddeley AD. Working memory. Oxford: Oxford University Press; 1986.
19Neath I. Modeling the effects of irrelevant speech on memory. Psychon Bull Rev 2000;7:403-23.
20Cowan N. Attention and memory: An integrated framework. Oxford: Oxford University Press; 1995.
21Baddeley AD, Hitch GJ. Working memory. In: Bower GA, editor. Recent Advances in Learning and Motivation Vol. 8. New York: Academic Press; 1974. p. 47-89.
22Jones DM, Macken WJ. Irrelevant tones produce an irrelevant speech effect: Implications for phonological coding in working memory. J Exp Psychol Learn Mem Cogn 1993;19:369-81.
23Sokolov EN. Perception and the conditioned reflex. Part Three Perception of Signal Stimuli. London: Pergamon Press; 1963. p. 161-258.
24Imbo I, Vandierendonck A. The role of phonological and executive working memory resources in simple arithmetic strategies. Eur J Cogn Psychol 2007;19:910-33.
25de Rammelaere S, Stuyven E, Vandierendonck A. The contribution of working memory resources in the verification of simple mental arithmetic sums. Psychol Res 1999;62:72-7.
26de Rammelaere S, Vandierendonck A. Are executive processes used to solve simple mental arithmetic production tasks? Curr Psychol Lett Behav Brain Cogn 2001;2:79-89.
27Hecht SA. Counting on working memory in simple arithmetic when counting is used for problem solving. Mem Cognit 2002;30:447-55.
28Logie RH, Baddeley AD. Cognitive processes in counting. J Exp Psychol Learn Mem Cogn 1987;13:310-26.
29Perham N, Macpherson S. Mental arithmetic and irrelevant auditory number similarity disruption. Irish J Psychol 2012. p. 181-92.
30Perham N, Banbury S, Jones DM. Auditory distraction impairs analytical reasoning performance. In: Katsikitis M. editor. Past Reflections, Future Directions: Proceedings of the 40 th Australian Psychological Society Annual Conference. Melbourne: APS; 2005. p. 238-42.
31Waldron SM, Patrick J, Morgan PL, King S. Influencing cognitive strategy by manipulating information access. Comput J 2007;694-702.
32Marsh JE, Hughes RW, Jones DM. Auditory distraction in semantic memory: A process-based approach. Journal of Memory and Language 2008;58:682-700.
33Bregman AS. Auditory scene analysis: The perceptual organisation of sound. Cambridge, MA: MIT Press; 1990.
34Macken WJ, Tremblay S, Houghton RJ, Nicholls AP, Jones DM. Does auditory streaming require attention? Evidence from attentional selectivity in short-term memory. J Exp Psychol Hum Percept Perform 2003;29:43-51.
35Marsh JE, Hughes RW, Jones DM. Interference by process, not content, determines semantic auditory distraction. Cognition 2009;110:23-38.
36LeCompte DC, Neely CB, Wilson JR. Irrelevant speech and irrelevant tones: The relative importance of speech to the irrelevant speech effect. J Exp Psychol Learn Mem Cogn 1997;23:472-83.
37Perham N, Marsh JE, Jones DM. Syntax and serial recall: How language supports short-term memory for order. Q J Exp Psychol (Hove) 2009;62:1285-93.
38Perham N, Whelpley C, Hodgetts H. Impaired memory for syntactical information in poor readers. Memory 2012.
39Martin RC, Wogalter MS, Forlano JG. Reading comprehension in the presence of unattended speech and music. J Mem Lang 1988:27:382-98.