Noise Health Home 

[Download PDF]
Year : 2010  |  Volume : 12  |  Issue : 49  |  Page : 210--216

Cross-modal distraction by background speech: What role for meaning?

John E Marsh, Dylan M Jones 
 School of Psychology, Cardiff University, Cardiff, United Kingdom

Correspondence Address:
John E Marsh
School of Psychology,Cardiff University, PO Box 901,Cardiff, CF10 3AT
United Kingdom


Mental tasks are susceptible to disruption by concurrent to-be-ignored speech. The goal of the present paper is to examine whether a theoretical framework successfully applied to irrelevant speech effects in serial recall-interference by process-can be extended to verbal tasks in which meaning is the basis of retrieval and to which the irrelevant sound is related to different degrees by meaning. That the semantic characteristics of the to-be-ignored sound interact with the predominance of semantic retrieval in the focal task to determine the degree of disruption is demonstrated in three settings: free recall, category-clustering and fluency. Source monitoring-the difficulty in discriminating episodic information on the basis of the sense modality (visual or auditory) in which it was presented-contributes in part to the disruption by speech. The power of alternative accounts-interference-by-content and attentional capture-to predict these outcomes is also discussed.

How to cite this article:
Marsh JE, Jones DM. Cross-modal distraction by background speech: What role for meaning?.Noise Health 2010;12:210-216

How to cite this URL:
Marsh JE, Jones DM. Cross-modal distraction by background speech: What role for meaning?. Noise Health [serial online] 2010 [cited 2022 Jul 6 ];12:210-216
Available from:

Full Text


Both personal experience and systematic study attest to the disruption of cognition by unwanted sound. Although hearing confers great benefit as a medium through which esthetic, social and intellectual life can be enriched, its sentinel character often brings unwelcome distraction. Part of this sentinel quality arises from what seems to be obligatory processing of sound, a feature most likely fashioned in the evolutionary development of sensory systems that capitalized on the auditory modality's omidirectionality, its capacity to detect remote events, as well as its availability in darkness or reduced visibility. Even in contemporary urban society, this quality still has utility. There are a great many settings in which it is beneficial to have attention drawn to auditory events (e.g., an auditory alarm or passing traffic) at the expense of processing of a task on which attention has been deliberately focused. However, interest in the present article centers on the inevitable price to be paid for being the sentinel of the senses. [1]

Most of what is understood about the distracting effects of sound on cognitive activity comes from the study of serial short-term memory, such as that involved in remembering a telephone number for the short-term. [2],[3] The damaging effect of irrelevant auditory information on cognitive activity in this serial short-term memory setting occurs even if the sound is relatively quiet [4] and when the focal, primary task contains no auditory component (with no possibility of auditory masking). In this context, these findings sit neatly within an interference-by-process account. Pre-attentive processing of the sound (processing without the engagement of attention or awareness) yields order cues, information which in turn conflicts with the deliberate seriation (or ordering) process involved in the serial recall task. [5] Although the focus on serial recall has been productive, theoretically it does not speak to the generality of distraction effects and indeed, suggests that their compass may be rather narrow. Slowly, evidence has been accumulating that semantic memory may also be vulnerable to distraction. Recently, there has been renewed interest in this work and this is what we attempt to review here.

 The Semanticity of Sound is Unimportant in Serial Recall

The meaning, or rather the semanticity, of irrelevant speech plays little, if any, role in the serial recall setting. For example, the meaningfulness of irrelevant speech (e.g., when it contains words in a language a participant understands) does not exacerbate the disruption to serial recall: narrative speech presented in a participant's native language produces comparable disruption to that produced by reversed narrative speech (largely incomprehensible), and speech that is presented in a language that the participant does not understand. [6] In addition, between-sequence semantic similarity or rather the similarity in the semantic content between to-be-remembered and to-be-ignored items is also an inconsequential factor. For example, Buchner, Irmen, and Erdfelder [7] found that serial recall of lists of two-digit to-be-remembered numbers was no more disrupted by to-be-ignored two-digit numbers (that were not part of the target list) than it was by non-words that comprised the phonemes of the numbers rearranged, or word combinations of the phonemes that were similar to those of the to-be-remembered numbers.

However, recently emerging evidence suggests that meaningful irrelevant speech does produce disruption in tasks in which meaning is used as the basis for retrieval. For example, Jones et al, [6] found that the meaning of sound had an impact on proof reading. Moreover, reading comprehension, involving the extraction and retention of propositions from prose, is also disrupted by meaningful speech. [8,9] Martin et al, [8] for example, found that comprehension of text passages, following an interpolated visual search task, was disrupted if the passages had previously been read in the presence of a spoken prose passage (in English) or random English words compared to white noise, instrumental music and random tones. In this study, sequences of randomly presented meaningful words were more disruptive than phonologically matched non-words. Although Martin et al. [8] report some disruption from meaningless sound (speech in a language foreign to the participant, and non-words) compared to quiet, these meaningless speech conditions did not differ from a broadband noise condition that has little or no acoustic variation; this suggests that the effect of auditory distraction on comprehension is qualitatively different from the case with serial recall. [6] That the focal, reading task involves the retention of propositions and not just verbatim recall (or serial order retention) appears a necessary pre-requisite for an effect of meaningfulness to be found. Tasks that can be accomplished by simply recalling a passage of text verbatim with less "weighting" on comprehension, fail to show an effect of meaning; in this setting meaningful and meaningless speech are equally damaging to performance compared to quiet. [10],[11]

The common denominator of focal tasks that demonstrate effects of meaning is that they call upon semantic processing. An effect attributable to the meaningfulness of irrelevant speech in the context of serial recall may not ordinarily emerge because typically items presented for serial recall (usually digits or consonants) are rather arbitrary and lack an appreciable semantic content. The to-be-remembered sequence therefore lacks syntax or semantic content that would help promote semantic processing, so that in serial recall the sequence is retrieved using subvocal repetition. [12]

 Semantic Free Recall

The focus of the current review is on retrieval of semantic information-from both lists and internally generated information-and its susceptibility to disruption from irrelevant speech. Despite the compelling evidence against between-stream semantic similarity as an influential agent in disrupting visual-verbal serial recall, [7] two studies have reported between-stream semantic similarity effects on tasks requiring free recall of semantic information, albeit with slightly different recall tasks. [13],[14] Fewer items were recalled from visually presented 16-item lists comprising examples drawn from a single semantic category (e.g., Birds: "cuckoo", "finch"), when dominant to-be-ignored auditory items of the same, as opposed to different, semantic category (e.g., "robin" and "eagle" as compared to "apple" and "banana") are presented in a retention interval prior to recall. Additionally, there is a tendency for semantically related irrelevant exemplars to be produced; [13] "Intrusion errors" is the term used to describe these productions of irrelevant items.

These findings appear consonant with the notion that the semantic effects are task-specific; they only occur if the focal task involves semantic processing, but other interpretations are possible. Indeed, it may be that the results are a mixture of semantic effects and episodic effects (one such candidate is the role of sensory modality as a cue to distinguish to-be-remembered from to-be-ignored events: source-monitoring). In the remainder of the article, we describe and review a series of our experiments [15],[16],[17] and attempt to fit them into an interference-by-process framework whilst at the same time excluding other possible interpretations. We approach this goal by posing three key questions: Is disruption by meaning a purely semantic effect? To what extent is a breakdown of source-monitoring responsible for semantic auditory distraction? Is retrieval accuracy determined by the relative dominance of the to-be-retrieved and to-be-ignored items?

Is Disruption by Meaning a Purely Semantic Effect?

No task is process pure. That is, nominally "semantic" tasks may contain several elements, some that are additional and functionally distinct from the semantic retrieval element. For example, the task requiring free recall of semantic information (e.g., 16-item lists of items drawn from a semantic category, say, Fruit [13],[14] ) has both a semantic component (knowledge of the target category and, hence, candidate responses) and an episodic memory component (the spatial and/or temporal context in which memories were experienced, for example, familiarity with the category items as ones belonging to the study episode). It seems possible that disruption by meaning of the irrelevant speech could have, at least in part, arisen from the episodic component of the semantic free recall task.

The tactic adopted to tease apart the relative contribution of semantic and episodic components was to use an array of tasks, each requiring recall of lexical-semantic information and embodying semantic memory, but each making differential demands on episodic memory. One task was free recall of list items drawn from single semantic categories (e.g., Tools); another involved a list of items drawn from several semantic categories presented randomly with the requirement to recall them by category (category clustering). A further two tasks involved lexical-generation from long-term memory, one requiring generation of items that belong to a semantic category (e.g., Birds: "eagle", "robin", "sparrow") known as semantic fluency, and one requiring generation of items that begin with a specific letter (e.g., f: "forest", "food", "fought") known as phonemic fluency. For all experiments, our participants were aged between 18 and 29 years and were undergraduate psychology students at Cardiff University. Approximately, two-thirds of the participants in each experiment were female. We now turn to consider each task in detail.

Semantic free recall

Three irrelevant speech conditions were used in the semantic free recall setting: to-be-ignored items related to the to-be-recalled items; to-be-ignored items unrelated to the to-be-recalled items and quiet. [15] In addition to this between-sequence semantic similarity manipulation, one group of participants was required to recall the list items in any order (free recall) whilst another group was required to recall the list items in the order they had occurred during list presentation (serial order). We reasoned that the results would be particularly supportive for a semantic interference-by-process if the disruption produced by meaning-either through meaningfulness, or between-sequence semantic similarity-was task-process-sensitive, i.e., apparent only when participants adopt a free recall, or a semantic categorization process, and not a serial-ordering process, for remembering. [Figure 1] illustrates that the disruption by to-be-ignored words semantically similar to the to-be-remembered list words is indeed task-process specific occurring only when recall is free and not when it is serial. This, of course, contrasts with the classic irrelevant sound effect which occurs predominately when, for the same lists, the task is serial. [18] {Figure 1}

Category clustering

Two experiments were carried out with a category-clustering task. [16] The first included three sound conditions: meaningful speech (English narrative), meaningless speech (reversed English narrative), and quiet. The second included to-be-ignored items related to the to-be-remembered items, to-be-ignored items unrelated to the to-be-remembered items and quiet. One group of participants in each experiment was required to recall the to-be-remembered items by semantic category and another group of participants was required to recall them in serial order. [Figure 2] shows that an effect of meaning on the number of correct responses (meaningfulness, see panel A and between-sequence semantic similarity, see panel B) arises only when the task involves retrieval by semantic category. Semantic organization of the responses in the semantic categorization groups was also impaired by the meaning of irrelevant speech. Although the category-clustering task is one of episodic memory (e.g., remembering which items from a given category were presented), the categorization of responses requires access to semantic memory. [19] The finding that the semantic categorization process is disrupted by the semantic properties of the irrelevant speech offers prima facie evidence that retrieval from semantic memory is susceptible to disruption via the meaning of irrelevant speech. Consistent with the interference-by-process account, these findings suggest that the effect of meaning is a joint product of task semantic processing and the obligatory processing of the meaning of irrelevant speech.{Figure 2}


Tasks that involve recalling lists will inevitably have an episodic component. Semantic fluency tasks minimize episodic memory: in semantic fluency, there is no list to remember; rather, participants generate their own list of items that belong to a given category name (e.g., Kitchen Utensils). Two experiments were carried out using semantic fluency. The first included three sound conditions: meaningful speech, meaningless speech (reversed speech) and quiet. The second included to-be-ignored items that were drawn from a category associated with the target-category (e.g., Vegetables when the target-category was Fruit), items drawn from a category non-associated to the target category (e.g., Tools when the target category was Fruit) and quiet.

[Figure 3] shows the outcome of these two experiments; fewer members of a category were produced in response to a category name (e.g., Vegetables) when participants ignored meaningful speech, as compared with meaningless speech and quiet (see Panel A). Moreover, fewer responses were produced when to-be-ignored items were drawn from a category associated to the target category as compared with those belonging to a different category. That semantic-lexical retrieval (lexical retrieval with semantic criteria), and not phonemic-lexical retrieval per se (lexical retrieval with phonemic criteria), was disrupted by the meaningfulness of the irrelevant speech was confirmed in a separate experiment requiring a different type of fluency, namely phonemic fluency. Here, participants generate words that begin with a given letter (e.g., f). The task, therefore, makes similar demands to semantic fluency in that it involves retrieval from long-term memory but instead of relying on semantic activation and search, it involves an algorithmic search of the lexicon based on initial letter sounds. Using phonemic fluency, we found no effect of the meaningfulness of sound, an outcome which bolsters the view that the focal, primary task and the irrelevant speech must share a similar process (e.g., semantic) for an effect of meaning to be found.{Figure 3}

Collectively, these findings from the semantic free recall, category clustering and fluency tasks offer powerful convergent evidence of mutual interference between semantic retrieval processes. Stripping away the episodic processing component from the task reveals a convincing disruptive effect of meaning therefore supporting the interference-by-process account. However, the episodic component may be important in findings other than a reduction in recall performance. Accurate episodic memory for list items can not only facilitate correct recall but prevent false recall. If one cannot bring to mind any surface (episodic) detail (e.g., visual-orthgraphic information in the case of the visual-presentation of a to-be-recalled list) for a candidate response, they can decide not to emit the response. It is to false recall that we now turn.

To What Extent is Source-Monitoring Responsible for Semantic Auditory Distraction?

In semantic free recall, participants often erroneously include the to-be-ignored auditory items when they are semantically related to the visually presented to-be-remembered items (false recalls). [13] This suggests a type of source modality confusion (e.g., [20] ): apparent uncertainty over the item's origin, whether the item was visually or auditorily presented, or merely generated by thought. [21] In the context of semantic distraction effects, if familiarity is the criterion by which the reporting of an item is judged, the retrieval mechanism may not be able to distinguish with enough accuracy whether the original events were in the auditory or in the visual modality and accordingly, auditory events may be misattributed to visual presentation. Although source-monitoring confusion had been previously mentioned as a possible explanation for the increase of false memory in semantic free recall, [13] there was a lack of confirmatory evidence until a recent study provided it. [15] We reasoned that the source-monitoring approach predicts that the prevalence of intrusions should be a positive function of the degree of temporal contiguity between the to-be-remembered items and the categorically related to-be-ignored items; the further in time that to-be-ignored items are removed from the presentation period of to-be-remembered items, the less is their likelihood of being "bound" with the source characteristics of the visual list (i.e., the study episode). [22]

One way of studying this is by restricting exposure of irrelevant sound to functionally distinct stages of the semantic free recall task. Typically, the semantic free recall procedure involves three phases: study, whereby the to-be-remembered items are presented in quiet; retention, during which-depending on the sound condition-participants are presented with quiet or to-be-ignored items that are either semantically related or unrelated to the to-be-remembered items, and test, in which participants produce their responses in quiet. [13],[14] The presence of these three phases allowed us to manipulate the temporal contiguity of the to-be-remembered in relation to the to-be-ignored items by presenting the to-be-ignored items such that they either coincided with the study, retention or test phases of the task. [Figure 4] shows the result of this experiment which was generally clear cut.{Figure 4}

Intrusions corresponding to the to-be-ignored items when presented during study were more numerous than when they were presented during the retention and test periods. Furthermore, intrusions from to-be-ignored items when presented during retention were more numerous than when they were presented during test.

The overall picture of a significant degree of confusion between what is seen and what is heard was further confirmed by the fact that confidence judgments by participants of their responses showed a misplaced confidence in to-be-ignored items belonging to the visually presented, to-be-remembered list. False recall-although increasing with increasing temporal contiguity-did not co-vary with the number of items correctly recalled which suggests that there may be two dissociable processes underpinning disruption on the semantic free recall task: an episodic component (responsible for increased false recall) and a semantic component (responsible for reduced correct recall).

Further confirmatory evidence that source-monitoring confusion is the basis of the between-sequence semantic similarity effect on false recall has been found using participant populations that are particularly poor at source-monitoring, such as older adults [23] and participants with low working memory capacity. [13] These participants commit more intrusions than their controls (younger adults, and participants with high working memory capacity, respectively; Sφrqvist [24] ).

What Role is Played by Relative Dominance of Items?

From the point of view of the interference-by-process account, between-sequence semantic similarity should produce disruption only when the to-be-ignored items are highly representative of the target category (output dominance; e.g., "robin" for Birds). By definition, output-dominant items are easily retrieved and thus are likely to compete for conscious awareness during retrieval of the to-be-remembered items. It seems intuitively likely that irrelevant items that are representative of a target category are more likely to disrupt performance because they have to be prevented from being retrieved and must also be monitored with regard to their origin of presentation. To address this prediction, a semantic free recall task was used in which the to-be-remembered list items were moderately representative of the category from which they were drawn. Two groups of participants were studied. One group of participants was presented with three sound conditions: quiet, to-be-ignored items that were highly representative of a category unrelated to the to-be-remembered items and to-be-ignored items that were highly representative of the same category from which the to-be-remembered items were drawn. The other group of participants was presented with quiet, unrelated to-be-ignored items weakly related to a non-target category, and related to-be-ignored items weakly related to the target category [Figure 5].{Figure 5}

In line with the process-based account, between-sequence semantic similarity impaired correct recall only when the related to-be-ignored items were highly representative of the target category and hence when they were more likely to compete for retrieval.

Collectively, these results provide insight into the nature of the effects of irrelevant sound on semantic retrieval tasks. For example, in this setting the disruption produced by irrelevant sound is qualitatively distinct from that in serial recall in that it is produced almost exclusively by the lexical-semantic, not acoustic, properties of the irrelevant stimuli. This is revealing in that it rules out the notion that the acoustic and semantic properties of irrelevant sound in the context of semantic retrieval tasks combine additively to determine disruption (cf). [25] However, despite this apparent success of the interference-by-process approach in accounting for semantic auditory distraction, there are at least two alternative accounts: interference-by-content and attentional capture. Both the accounts fail to adequately account for the data with serial recall [5] but how do they fare in relation to semantic auditory distraction?

 Alternatives to Interference by Process


According to this approach, disruption occurs as a mere consequence of auditory stimuli gaining access to the same representational space (in memory) as the to-be-recalled items. Distinct from the interference-by-process account, the interference-by-content account supposes that distraction occurs as a direct and passive consequence of the structural similarity in the post-categorical identity of to-be-ignored and to-be-recalled items. At first blush, this account seems to be readily compatible with the findings from semantic auditory distraction: within working memory the semantic representations of the to-be-recalled items may be degraded (via mutual feature-overwriting between to-be-recalled and to-be-ignored items) or otherwise made less accessible as a function of their semantic similarity to the to-be-ignored items. [26]

However, a number of findings undermine the interference-by-content account. First, on this account, the mere presence of semantic content within the to-be-remembered and to-be-ignored material should not be sufficient to produce disruption. Thus, the semantic attributes of the speech should have disruptive potency regardless of whether the to-be-recalled items are retrieved according to their meaning (semantic free recall and semantic categorization) or serial order. This is because the semantic content of the speech should be represented in working memory regardless of the task instructions. [27] However, at odds with the interference-by-content account (but consistent with the interference-by-process account), is the finding that disruption fails to emerge if the task instructions demand serial recall. [15],[16] Therefore, it appears that a shift in the nature of the dominant process/strategy used to support performance in the focal task, not simply the semantic richness of the to-be-remembered items, is what renders tasks susceptible to disruption from the (apparently obligatory) processing of the semantic content of the irrelevant speech.

Second, the interference-by-content approach supposes that irrelevant verbal items presented during study or retention as compared with test should produce more disruption because of a greater likelihood of feature overwriting. However, in practice, comparable disruption is found regardless of whether the to-be-ignored items were presented at study, retention or test phases. [15] Third, the interference-by-content account assumes disruption only in settings in which working (or episodic) memory-and, in particular, temporary storage-is involved such as in free recall whereby a list of items has to be remembered. Unlike the interference-by-process account, the interference-by-content account, therefore, does not appear to readily predict the disruption found on semantic verbal fluency which requires direct access to semantic memory and for which no remembering of list items is required. [17] However, one could argue that working memory is involved in semantic fluency for the storage of responses prior to, or after, emission and that the reduction in fluency reflects interference with this form of memory. However, if this were true, forgetting of responses should occur and repetition errors should increase; this is not found.

Attentional capture

A variant of the capture-based account of distraction assumes that attention may be captured away from the focal task to the extent that the semantic features of the to-be-ignored items are primed by the to-be-remembered item. [28] This predicts an effect of between-sequence semantic similarity, regardless of whether the focal task requires semantic recall (free recall or semantic categorization) or serial recall. This is because automatic semantic priming occurs "full blown", regardless of the nature of the focal task. [29] That the between-sequence semantic similarity effect only occurs with semantic free recall and semantic categorization clearly weakens this approach. Moreover, that the meaning of speech does not produce disruption in tasks that do not require extensive semantic analysis (serial recall and phonemic fluency) and does produce disruption in those that do (category clustering and semantic fluency) rules out the idea that it is disruptive merely because it captures attention regardless of the nature of the focal task processes.


Semantic auditory distraction is a form of distraction that is in many ways distinct from that found with serial recall, but at heart both are served by a generic class of explanatory construct, namely interference-by-process. With semantic auditory distraction, the focal task processes involved are of course different from those in serial recall but the underpinnings of the effects may be the same: a clash of an automatic and a similar deliberate process that reduces task performance efficiency. In the context of serial recall, the automatic, pre-attentive process is the serial ordering of the changing acoustic elements of the irrelevant sound. The deliberate process is the ordering of the to-be-remembered items in the primary focal task via serial rehearsal. For semantic retrieval, the automatic spreading semantic activation from the irrelevant material impairs the deliberate process of using such semantic activation(s) for the purpose of retrieval. Moreover, the interference-by-process seems to be more generalizable than the interference-by-content and attentional capture accounts. Its constructs more readily predict that distraction will occur in complex tasks that require semantic retrieval such as comprehension, [8],[9] essay writing [30] and proof reading. [6] Regardless of the complexity of the focal, primary task, the two streams of information-that of the current cognitive activity and that of the irrelevant speech-do not come into conflict for what they contain but for how they are processed. Interventions for the purpose of minimizing auditory distraction must aim to reduce both the intelligibility of the potentially distracting background sound (reducing semantic processing) and its acoustic complexity (reducing serial processing).


The research reported in this article received financial support from an ESRC grant awarded to Dylan Jones and C. Philip Beaman (RES-062-23-1752).


1Hughes RW, Jones DM. Indispensable benefits and unavoidable costs of unattended sound for cognitive functioning. Noise Health 2003;6:63-76.
2Jones DM. The cognitive psychology of auditory distraction: The 1997 BPS Broadbent Lecture. Br J Psychol 1999;90:167-87.
3Jones DM, Hughes RW, Macken WJ. Auditory distraction and serial memory: The avoidable and the ineluctable. Noise Health 2010;12:201-9.
4Tremblay S, Jones DM. Change of intensity fails to produce an irrelevant sound effect: Implications for the representation of unattended sound. J Exp Psychol Hum Percept Perform 1999;25:1005-15.
5Jones DM, Tremblay S. Interference in memory by process or content? A reply to Neath (2000). Psychon Bull Rev 2000;7:550-8.
6Jones DM, Miles C, Page J. Disruption of proofreading by irrelevant speech: Effects of attention, arousal or memory? Appl Cogn Psychol 1990;4:89-108.
7Buchner A, Irmen L, Erdfelder E. On the irrelevance of semantic information for the "irrelevant speech" effect. Q J Exp Psychol 1996;49A:765-79.
8Martin RC, Wogalter MS, Forlano JG. Reading comprehension in the presence of unattended speech and music. J Mem Lang 1988;27:382-98.
9Oswald CJ, Tremblay S, Jones DM. Disruption of comprehension by the meaning of irrelevant sound. Memory 2000;8:345-50.
10Banbury S, Berry DC. Habituation and dishabituation to speech and office noise. J Exp Psychol Appl 1997;3:181-95.
11Banbury S, Berry DC. Disruption of office-related tasks by speech and office noise. Br J Psychol 1998;89:499-517.
12Perham N, Marsh JE, Jones, DM. Syntax and serial recall: How language supports short-term memory for order. Q J Exp Psychol 2009;62:1285-93.
13Beaman CP. The irrelevant sound effect revisited: What role for working memory capacity? J Exp Psychol Learn Mem Cogn 2004;30:1106-18.
14Neely CB, LeCompte DC. The importance of semantic similarity to the irrelevant speech effect. Mem Cognit 1999;27:37-44.
15Marsh JE, Hughes RW, Jones DM. Auditory distraction in semantic memory: A process-based approach. J Mem Lang 2008;58:682-700.
16Marsh JE, Hughes RW, Jones DM. Interference by process, not content, determines semantic auditory distraction. Cognition 2009;110:23-38.
17Marsh JE, Hodgetts HM, Hughes RW, Jones DM. Verbal fluency falters under distraction. 2010 [In Press].
18Beaman CP, Jones DM. Irrelevant sound disrupts order information in free as in serial recall. Q J Exp Psychol 1998;51A:615-36.
19Bousfield WA. The occurrence of clustering in the recall of randomly arranged associates. J Gen Psychol 1953;49:229-40.
20Henkel LA, Franklin N, Johnson MK. Crossmodal source-monitoring confusions between perceived and imagined events. J Exp Psychol Learn Mem Cogn 2000;26:321-35.
21Smith SM, Tindell DR, Pierce BH, Gilliland TR, Gerkens DR. Source memory failure in episodic confusion errors. J Exp Psychol Learn Mem Cogn 2001;27:362-74.
22Barnhardt TM, Choi H, Gerkens DR, Smith SM. Output Position and word relatedness effects in a DRM paradigm: Support for a dual-retrieval process theory of free recall and false memories. J Mem Lang 2006;55:213-31.
23Bell R, Buchner A, Mund I. Age-related differences in irrelevant-speech effects. Psychol Aging 2008;23:377-91.
24Sφrqvist P. The role of working memory capacity in auditory distraction: A review. Noise Health 2010; 12:217-24.
25Beaman CP. Auditory distraction from low-intensity noise: A review of the consequences for learning and workplace environments. Appl Cogn Psychol 2005;19:1041-64.
26Oberauer K, Lange E, Engle RW. Working memory capacity and resistance to interference. J Mem Lang 2004;51:80-96.
27Rouibah A, Tiberghien G, Lupker SJ. Phonological and semantic priming: Evidence for task-independent effects. Mem Cognit 1999;27:422-37.
28Cowan N. Attention and memory: An integrated framework. Oxford: Oxford University Press; 1995.
29Neely JH, Kahan TA. Is semantic activation automatic: A critical evaluation. In: Roediger HL, Nairne JS, Neath I, Surprenant AM, editors. The nature of remembering: Essays in honor of Robert G Crowder. Washington, DC: American Psychological Association; 2001. p. 69-93.
30Olive T. Working memory in writing: Empirical evidence from the dual task technique. Eur Psychol 2004;9:32-42.