The aim of the present study was to test which tasks are suitable for work in open-plan offices according to how susceptible they are to disruption produced by the mere presence of irrelevant speech. The tasks were chosen to tap fundamental capacities of office work involving: search for relevant information, remembering material, counting, and generation of words. The hypothesis was that tasks requiring semantic processing should be impaired by irrelevant speech. To determine the magnitude of performance decrease, two sound conditions (quiet, irrelevant speech) were compared. The results showed that tasks based on episodic short-term-memory and rehearsal of the presented material were more sensitive to disruption by irrelevant speech than tasks which did not require rehearsal or were based on long-term memory retrieval. The present study points to the inappropriateness of tasks, such as information search and remembering of material, for work environments within which irrelevant speech is ubiquitous.
Keywords: Long-term memory, noise, semantic auditory distraction, working memory
|How to cite this article:|
Jahncke H. Open-plan office noise: The susceptibility and suitability of different cognitive tasks for work in the presence of irrelevant speech. Noise Health 2012;14:315-20
|How to cite this URL:|
Jahncke H. Open-plan office noise: The susceptibility and suitability of different cognitive tasks for work in the presence of irrelevant speech. Noise Health [serial online] 2012 [cited 2020 Jul 13];14:315-20. Available from: http://www.noiseandhealth.org/text.asp?2012/14/61/315/104901
| Introduction|| |
There is a widespread concern for how undesired sounds in offices affect cognition and performance. , Many noise sources are potentially disruptive; however, several studies have shown that irrelevant speech from colleagues is the most troublesome, especially for performance and perceived disturbance in open-plan offices.  In addition, there are studies which indicate that some tasks are more affected by irrelevant speech than others. , The aim of this study is to further specify which tasks are more-as compared to less-sensitive to disruption via irrelevant speech. This may give an indication of which tasks are suitable or inappropriate to work with in open-plan offices as these designs promote noisy work environments.
Several studies have shown that reading comprehension, , information search,  recall of previously presented semantic information [8,9] and writing processes,  which all require semantic processing (i.e., processing of meaning), are detrimentally affected by irrelevant speech. This outcome is commonly explained as a cognitive conflict between the automatic semantic processing of irrelevant background speech and the simultaneous semantic processing of those tasks. , Support for the automatic semantic processing of speech sounds, even if the sound is unattended, comes from both neuroscientific  and behavioral  studies.
However, if a task only involves rehearsal of information in a specific order (e.g. recall of the digits in a telephone number) rather than semantic processes, other properties of the sound (i.e. sound that changes in terms of frequency or pitch) have been shown to be more disruptive than the semantic aspects of sound. In this case a processing conflict occurs when the order of events in the sound and of the focal task are simultaneously processed, which is called the irrelevant sound effect. Perham, Banbury, and Jones  have shown that retrieval instructions can influence the extent to which rehearsal is adopted and consequently the degree of disruption. Therefore, the mere presence of sound (e.g. irrelevant speech) can impair performance in tasks that require serial processing or where the most efficient strategy is rehearsal, as long as the sound changes across time. It has also been shown that counting, a fundamental work task for several professions, is impaired by irrelevant sounds, but this occurs regardless of whether the background noise consists of speech numbers.  Even if counting does require some processing of meaning, these results suggest there are few demands on semantic processing when the math task is simple.
Jahncke, Hongisto and Virjonen  have tested the magnitude of disruption by irrelevant speech on a variety of cognitive tasks. The results showed, for instance, that a word memory task was more impaired by speech with high levels of intelligibility than a math task was. However, as the two tasks contained different materials, it is difficult to tell whether it is the complexity, the difficulty level, or the cognitive processes that make tasks differ in sensitivity to disruption from irrelevant speech. Furthermore, Marsh and Jones  have investigated the effects of meaningful speech on one task with different semantic demands. They showed that a word generation task with semantic retrieval (semantic fluency, such as generating members of a specific semantic category) was disrupted by meaningful background speech but phonemic retrieval (phonemic fluency, such as generating words that all begin with the same phoneme) was not. Moreover, the disruption increased when the background speech was semantically related to the semantic fluency task (e.g., when the spoken words were members of the same semantic category as the to-be-retrieved words). The main focus in the present study is to test which tasks (the components of which are relevant to office work), are more or less susceptible to disruption from background speech. The first hypothesis is that performance in tasks based on semantic processing (e.g., word memory, semantic fluency and information search) will be more impaired by irrelevant background speech than tasks based on other or less semantic cognitive processes (e.g., counting and phonemic fluency). A related aim is to test whether changes in types of processing demands modulate the task's sensitivity to disruption from irrelevant speech. For this purpose, a novel task (word memory: category- and structural word finding; further described under method section) was developed, which changes the requirements of semantic processing by varying the processing instruction (i.e., changing the mental operations that participants perform on the to-be-remembered material), without changing the test material per se. Research on incidental learning has, for instance, shown that manipulating instructions in this way indeed changed the processing requirements.  More specifically, the second hypothesis is that performance in the novel task will be more impaired when the participants are following the semantic instruction compared to the structural instruction.
| Methods|| |
The experiment was carried out in sound proof rooms. A total of 24 persons recruited from the University (mean age = 25; 8 females) took part in the study in exchange for two cinema tickets. All participants reported their hearing and vision to be normal.
Two noise conditions were compared: quiet versus irrelevant speech. The speech was presented by individual laptops and headphones. The irrelevant speech was played back at an average level of 51 dB (A), which is a typical overall level in open-plan offices.  The quiet condition was approximately 30 dB (A).
The original speech was recorded in an anechoic room and should not be overvalued as a realistic office noise. However, the irrelevant speech conveyed meaningful information and consisted of 120 different sentences spoken by a female actress and 120 different sentences spoken by a male actor, thus yielding 240 sentences. Each sentence had simple semantic content, e.g., "Japanese cars are cheap." The duration of each sentence was three seconds. The sentences were spoken in Swedish with clear articulation. The sentences were looped in random order during the experimental session.
Memory: Category- and structural word finding
The aim in relation to these tasks was to manipulate the cognitive processes involved while performing the tasks. Therefore, two versions of the instructions were used but the material was kept the same (i.e., lists of words). In both cases three lists with ten words each were presented before the participants had to recall the words in free order. In the first version of the memory task the participants were instructed to remember all words within a named category (e.g., a toy) while searching lists of words. This part of the task was intended to measure memory for semantic targets. In the second version of the memory task the participants were instructed to remember words beginning with a certain letter. This version of the task was intended to measure memory for structural targets. Even if these tasks are based on either semantic or structural analysis of the targets, both tasks put high demands on episodic memory and presumably rehearsal of the to-be-remembered material. Both versions of the task are relevant to office work because they tap processes required to search through material and memorize relevant parts. The words were only included once (i.e., never repeated) to avoid recognition or interference with earlier presented words. Each list with words (shown in rows) was visible for 6 seconds. There were ten trials (in total 30 lists) for each sound condition and task. For each trial (three lists) six words were correct. The correct words were divided between the lists in two ways. In every second trial there were two correct words per list and in the other half, there were one correct word in the first list, three correct words in the second list, and two correct words in the last list.
For the category memory task, the same numbers of categories were selected to both versions of the task according to Overschelde, Rawson and Dunlosky×s  difficulty level rankings. The following categories were included within one of the versions: a unit of time, a type of footwear, a tree, a weapon, a type of dance, a type of reading material, a type of fabric, a toy, a flower and a herb. In the other version the following categories were included: a sport, a vegetable, a substance for flavoring food, an alcoholic beverage, a type of human dwelling, a thing that flies, a liquid, a carpenter×s tool, a thing that is green and a part of speech. The two versions were counter balanced between participants over the two sound conditions.
For the structural memory task the participants were instructed to remember words beginning with the following Swedish letters in one version of the task: H, N, A, V, O, J, T, Ä and L; and in the other version: E, Å, M, R, Ö, I, Y, P, U and K. The two versions were counter balanced between participants over the two sound conditions.
Word generation: Semantic- and phonemic fluency
Two versions of the word generation task  were developed by Jahncke et al.,  and adopted in the present study to test semantic- and phonemic fluency. The procedure for the Semantic fluency task was that the participants were given 60 seconds to generate (write) as many exemplars from a given semantic category as possible. Directly after the first category, another category was presented with the same time limit. Jahncke et al.,  used the updated and expanded version of Battig and Montague category norms developed by Overschelde, Rawson and Dunlosky  for category selection, to keep the difficulty level fairly constant between the versions of the task. The categories were: A four-footed animal, a thing that makes noise, a musical instrument and a kitchen utensil. The scores were the sum of correctly generated words and the task took two minutes to perform (i.e., two categories).
In the Phonemic fluency task the participants were presented with a letter and told to generate as many words as possible during 60 seconds that begin with the given letter. However, they were not allowed to generate variants of the same word: e.g. sail, sailing, and they were instructed to avoid producing proper nouns (e.g. names of places, people and products). Jahncke et al. based the letter selection to the phonemic fluency task on the study by Borkowski, Benton and Spreen,  which defined the difficulty level of the letters according to the associated word frequency. In each of the six versions of the task two letters were presented. One letter were taken from the easy range category of difficulty (i.e., H, D, M, A, B, F) and one from the moderate range category (i.e., O, N, E, G, L, R), to keep the difficulty level fairly constant between the versions of the task. The scores were the sum of correctly generated words and the task was given two minutes to perform. Both of the fluency tasks require retrieval from long-term memory, with either a semantic or phonemic activation of words. These tasks do not involve rehearsal of a to-be-remembered material and are relevant to office work in the respect that they tap processes to generate written texts and ideas.
Information search task
The information search task was adopted from Jahncke and Halin.  This task was originally designed to have relevance for office work because it taps key processes required to search through and comprehend the contents of a table of information, while successively updating and memorizing which information is most correct according to a target criterion. This task is therefore based on episodic memory and rehearsal and updating of the to-be-remembered material.
In the task seven columns of information were given about price, location, area, year etc. The twenty rows kept together the information about a given object (e.g., a person, a house or a country). The participants were asked to find the object that met a set of criteria, by using three columns. Each experimental block consisted of twelve questions and the time was limited to one minute per question before a new question was presented. The scores were the sum of correct answers and the average answer time. The task took 12 minutes to perform.
This task was an addition task where participants were instructed to add triple-digit numbers (e.g., 146 + 309).  This task was included as arithmetic is relevant to office work, especially for economists, engineers etc. Each experimental block consisted of 25 pairs of triple-digit expressions presented one by one together with an answer box. Every second expression had two carry over operations and the others had one carry over operation. Fürst and Hitch  have argued that this carrying component in mental arithmetic puts high demands on inhibition and memory processes of interim results in working memory. This task therefore require some kind of episodic or working memory (i.e., storage of a solution), though presumably with less episodic demands than the word memory tasks and fewer semantic demands than the fluency tasks. In each of the six versions of this task there were different numbers to sum up. The answer time was set to 18 seconds per expression. The scores were the sum of correct answers and the average answer time. The task took approximately 7 minutes to perform.
The hypotheses were tested by using a within-person design with two noise conditions (quiet; irrelevant background speech) and four different tasks expected to involve diverse cognitive processes (word generation; memory from semantic-/structural word finding; counting; and information searching).
Data collection took place in sound proof rooms. The participants sat in isolation and worked with the cognitive tasks on a laptop and they had headphones on during the whole session. First the participants undertook a short practice session of all tasks in quiet. Then they proceeded through the other tasks and noise conditions in random order. The following restrictions for the randomization were used: the same task was not allowed to be performed in noise and quiet directly after each other; and a maximum of three tasks with the same noise condition were allowed to be presented in a row. In this manner each participant performed all of the tasks during both quiet and noise but in a random order. The session took about one hour to complete.
| Results|| |
The analysis revealed a main effect of background speech on information search, F(1, 23) =9.47, P < 0.005, partial ŋ 2 = 0.29; on memory from category word finding, F(1, 22) =7.12, P < 0.02, partial ŋ 2 =0.25; and on memory from structural word finding, F(1, 23) =14.14, P < 0.001, partial ŋ 2 = 0.38; indicating lower performance with background speech compared to quiet in line with the hypothesis [Figure 1]. More specific, performance decreased from quiet to background speech with 8.7% for the information search, 6.5% for memory from category word finding, and 10.6 % for memory from structural word finding.
|Figure 1: Percentage of correct answers in the cognitive tasks during quiet and irrelevant speech|
Click here to view
Further, in line with the hypothesis there was no effect of background speech on math performance, F(1, 23) =0.17, P = 0.68, partial ŋ 2 = 0.01; and on phonemic fluency, F(1, 23) =1.06, P =0.31, partial ŋ 2 = 0.04. Unexpectedly there was no effect of background speech on semantic fluency, F(1, 23) =0.26, P =0.62, partial ŋ 2 = 0.01. Performance decreased from quiet to background speech with 3.3% for math, 4.2% for phonemic fluency and 2.4% for semantic fluency. No effects of background speech were found on answer time in the information search and math task.
| Discussion|| |
This experimental study addressed the issue of which tasks (the components of which are relevant to office work), are more or less susceptible to disruption by the mere presence of irrelevant speech.
The main findings are that the information search task and the two memory based tasks (structural- and category words finding) were affected by irrelevant speech, but the math task and the two fluency tasks (semantic- and phonemic fluency) were not. This was not entirely expected as it was hypothesized that the semantic fluency task would be sensitive to irrelevant speech because of the requirements of semantic processing and the presence of semantic information in the irrelevant speech. Furthermore, it was hypothesized that the structural-memory task would be resistant to background speech as it arguably does not involve semantic processing.
First, to understand why both of the memory based tasks were impaired by background speech, and neither of the two fluency tasks, it is important to note the differences in the functional attributes of these tasks. The two memory tasks were based on episodic memory and presumably rehearsal of the to-be-remembered material which is the main strategy for keeping information available for report, and hence, should make the tasks susceptible to disruption from changing-state characteristics of the sound. In contrast, the fluency tasks required retrieval of words from long-term memory and were accordingly not based on rehearsal of a to-be-remembered material. Jones, Marsh and Hughes  argued in a recent study that different stages of a task (i.e. encoding, rehearsal and retrieval) can be impaired by irrelevant speech. They showed that semantic fluency, which is based on retrieval, is only vulnerable to speech if the speech is meaningful; and is even more impaired if the speech is related semantically to the retrieval category. However, meaningful speech was used in the current experimental study and semantic fluency appeared invulnerable to disruption. Although speculative, the difference between Jones et al.,  and the current findings may depend on the level of abstractness of the categories selected, the size of those categories, or on the semantic properties of the background speech. This clearly needs to be subject of further investigation.
The remaining results may also be explained within the perspective of different memory processes and stages. Another variable of interest is the presence of visual supports which will require different levels of rehearsal. Tasks were presented visually, but some provided timed presentations while others maintained more visual support. Indeed with full visual support in timetable-type tasks, the disruptive effect of irrelevant speech is eliminated.  In the context of the information search task used in the current study, episodic memory and rehearsal is necessary because the task required the participants to search through a table of information whilst remembering (and updating) which target was closest to the given statement. This use of rehearsal presumably made the task vulnerable to disruption via background speech. The math task, on the other hand, only included a simple addition of numbers, and perhaps did not require as much by way of rehearsal as each operation was visible during the whole answer time and the answer did not need to be memorized or updated, but could be written directly in the answer box. The low demands on rehearsal in the simple math task might therefore explain its resistance to any disruption produced by the background speech. The present results suggest that irrelevant speech produces poorer performance in tasks with high episodic memory and rehearsal demands (i.e. word memory and search task), but very little reduction in performance on tasks that make fewer demands on episodic memory processes (i.e. math), or those tasks based on long term memory (i.e. fluency task). This is in line with earlier studies showing that episodic memory tasks are more disrupted by noise than are semantic tasks  and that the retrieval strategy can influence the degree of disruption by noise  However, as these tasks are complex and there are different mnemonic properties (in terms of processes and content) involved across the tasks that have been compared, it is difficult to draw strong conclusions. Some of the limitations of this type of research stem not necessarily from flaws in design, but simply from the noteworthy attempt to describe the influences of distraction on completion of complex tasks.
Further, it can be questioned from another point of view why the two word memory tasks did not differ as hypothesized. The second hypothesis stated that performance should be more impaired when the participants were following the semantic instruction compared to the structural instruction. A plausible explanation is that the initial part of the tasks, requiring analysis of either semantic or structural targets, was not sensitive to irrelevant speech, though the following processes were (i.e. episodic memory and rehearsal). As both tasks required the latter processes no significant differences between the tasks was obvious.
However, it is important to note that the absence of meaningless irrelevant speech as a control condition potentially obscures identification of the variable(s) responsible for producing the irrelevant speech effect in the context of the word memory tasks: without a meaningless irrelevant speech condition, it cannot be ascertained whether the tasks were impaired by the changing acoustical properties of the irrelevant sound such as in serial recall (in which one would expect no additional disruption from meaningful speech compared to meaningless speech), or whether the disruption was a semantic irrelevant sound effect (in which one would expect additional disruption from meaningful speech).  Also, Perham et al.,  indicated that minimizing the demands on rehearsal in a task made the irrelevant speech effect disappear, compared to when the demands on rehearsal were high.
Furthermore, the irrelevant speech was composed of randomly presented sentences - not discourse. While the sentences contained semantic information, they were not related to one another. Speech distraction in an office situation might be more discourse oriented (overheard conversations, etc.) and as such a comparison between the random sentences and connected discourse might be more meaningful than comparisons with meaningless speech and random sentences, if the purpose is to discuss effects on office performance. However, in this study we focused on random sentences as it can also be distracting to overhear half of a conversation (i.e. telephone calls from colleagues) where the heard sentences become more random for the listener.  As mentioned earlier, the present study did not aim at ruling out theoretical explanations of why background speech is disruptive. The main focus in this study was merely to explore which tasks are suitable to work with in open-plan offices where exposure to irrelevant speech sounds is frequent. The ecological validity of the tasks included can therefore be discussed. The tasks investigated were chosen to tap fundamental characteristics of office work, involving the search for relevant information, remembering material, counting, and generating words (which underpins language production). Further research is needed to investigate a greater variety of open-plan office settings (i.e. speech surroundings) and tasks with different attributes. As researchers, we are left with the difficult task of attempting to isolate components of tasks that are susceptible to distraction, whilst concluding as the cause of distraction on integrative tasks that embody more than one of the identified component processes. This becomes problematic when one wishes to make important recommendations about office environments based on the results. Functional and practical conclusions from this study are necessarily limited as a result.
In summary, some basic conclusions can be drawn. The automatic processing of irrelevant speech impedes the cognitive processes required to rehearse to-be-remembered material. According to the results of this study it appears that tasks which require episodic memory and the process of rehearsal-such as information searching and word memory-are better performed during quiet. Word fluency (based on material stored in long-term-memory) and simple counting however, are less susceptible to disruption via the presence of irrelevant speech. Since work tasks differ in how vulnerable they are for background speech, it is important to consider which work environments are optimal for each kind of task component.
| References|| |
|1.||Szalma JL, Hancock PA. Noise effects on human performance: A meta-analytic synthesis. Psychol Bull 2011;137:682-707. |
|2.||Jahncke H, Hygge S, Halin N, Green AM, Dimberg K. Open-plan office noise: Cognitive performance and restoration. J Environ Psychol 2011;31:373-82. |
|3.||Kaarlela-Tuomaala A, Helenius R, Keskinen E, Hongisto V. Effects of acoustic environment on work in private office rooms and open-plan offices-longitudinal study during relocation. Ergonomics 2009;52:1423- 44. |
|4.||Beaman 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. |
|5.||Jahncke H, Hongisto V, Virjonen P. Cognitive performance during irrelevant speech: Effects of speech intelligibility and semantic difficulty. Applied Acoustics 2012; Accepted from: 10.1016/j.apacoust.2012.08.007. |
|6.||Oswald CJP, Tremblay S, Jones DM. Disruption of comprehension by the meaning of irrelevant sound. Memory 2000;8:345-50. |
|7.||Sörqvist P, Halin N, Hygge S. Individual differences in susceptibility to the effects of speech on reading comprehension. Appl Cogn Psychol 2010;24:67-76. |
|8.||Bell R, Mund I, Buchner A. Disruption of short-term memory by distractor speech: Does content matter? Q J Exp Psychol (Hove) 2011;64:146-68. |
|9.||Beaman CP. The irrelevant sound phenomenon revisited: What role for working memory capacity? J Exp Psychol Learn Mem Cogn 2004;30:1106-18. |
|10.||Sörqvist P, Nöstl A, Halin N. Disruption of writing processes by the semanticity of background speech. Scand J Psychol 2012;53:97-102. |
|11.||Marsh JE, Hughes RW, Jones DM. Auditory distraction in semantic memory: A process-based approach. J Mem Lang 2008;58:682-700. |
|12.||Marsh JE, Jones DM. Cross-modal distraction by background speech: What role for meaning? Noise Health 2010;12:210-6. |
|13.||Beaman CP, Bridges AM, Scott SK. From dichotic listening to the irrelevant sound effect: A behavioral and neuro imaging analysis of the processing of unattended speech. Cortex 2007;43:124-34. |
|14.||Sörqvist P, Marsh JE, Jahncke H. Hemispheric asymmetries in auditory distraction. Brain Cogn 2010;74:79-87. |
|15.||Perham N, Marsh JE, Jones DM. Syntax and serial recall: How language support short-term memory for order. Q J Exp Psychol 2009;62:1285-91. |
|16.||Perham N, Banbury SP, Jones DM. Reduction in auditory distraction by retrieval strategy. Memory 2007;15:465-73. |
|17.||Buchner A, Steffens MC, Irmen L, Wender KF. Irrelevant auditory material affects counting. J Exp Psychol Learn Mem Cogn 1998;24:48-67. |
|18.||Shulman HG. Similarity effects in short-term memory. Psychol Bull 1971;75:399-415. |
|19.||Van Overschelde JP, Rawson KA, Dunlosky J. Category norms: An updated and expanded version of the Battig and Montague (1969) norms. J Mem Lang 2004;50:289-335. |
|20.||Troyer AK, Moscovitch M, Winocur G, Alexander MP, Stuss D. Clustering and switching on verbal fluency: The effects of focal frontal- and temporal-lobe lesions. Neuropsychologia 1998;36:499-504. |
|21.||Borkowski JG, Benton AL, Spreen O. Word fluency and brain damage. Neuropsychologia 1967;5:135-40. |
|22.||Jahncke H, Halin N. Performance, fatigue and stress in open-plan offices: The effects of noise and restoration on hearing impaired and normal hearing individuals. Noise and Health 2012;14:260-72. |
|23.||Fürst AJ, Hitch GJ. Separate roles for executive and phonological components of working memory in mental arithmetic. Mem Cognit 2000;28:774-82. |
|24.||Jones DM, Marsh JE, Hughes RW. Retrieval from memory: Vulnerable or inviolable? J Exp Psychol Learn Mem Cogn 2012;38:905-22. |
|25.||Perham N, Banbury S. The role of rehearsal in a novel call centre-type task. Noise Health 2012;14:1-5. |
|26.||Boman E, Enmarker I, Hygge S. Relative strength of noise effects on memory processes in pupils and teachers depending on noise source and age. Noise Health 2005;7:11-26. |
|27.||Marsh JE, Hughes RW, Jones DM. Interference by process, not content, determines semantic auditory distraction. Cognition 2009;110:23-38. |
|28.||Monk A, Fellas E, Ley E. Hearing only one side of normal and mobile phone conversations. Behav Inf Technol 2004;23:301-5. |
Environmental Psychology, Faculty of Engineering and Sustainable Development, University of Gävle, SE-801 76 Gävle
Source of Support: None, Conflict of Interest: None