Miss Emma Elizabeth Dorothy Meredith
What does the dichotic listening task tell us about how we attend to information? Discuss this with reference to early and late selection models of attention.
Attention is a selection process and is invaluable to our everyday lives. We use our attention for many things such as crossing the road, watching over our children, or even while multi-tasking (which is an invaluable skill in today’s world). Psychologist William James (890), talked about two modes of attention; Active attention which is indicative of an individual’s goals and expectations and controlled in a top-down way; and passive attention which is controlled in a bottom-up way and is effected by external stimuli (Eysenck, Keane, 2010). The way attention works brings about a host of questions which has generated much research in to this field; questions such as how much information can we take in at once? What happens to the information to which we are not attending? Do distractions from unwanted information cause attention to fail? Other questions are how do we process the information coming through our senses? And when in the process do we filter out that which is not relevant? (Naish, 2010). This essay will be concerned with where the filter lies in the process and will compare and contrast early and late models of selection, drawing on research in this area which has used dichotic listening tasks to help explain how we attend to information.
The dichotic listening task is a method commonly used by researchers in psychology to explore selective attention. Participants are asked to wear headphones, in which they are then presented with multiple sounds representing that of the real world. These different sounds are played in each ear and are easily manipulated (Naish, 2010). Broadbent (1952, 1954) used the dichotic listening task to support the bottleneck theory of attention. This theory proposed that information is selected on the basis of ‘low level’ physical characteristics such as location of sound, pitch and gender of speaker, and only this information is processed which would suggests that this bottleneck in the attentional system is such that only a limited amount of sensory information passes through it. Broadbent argued that this bottleneck occurs early on in the filtering process. To demonstrate this theory, Broadbent used the dichotic listening task in which he used two conditions where the participants are asked to listen and repeat different numbers. Pairs of different numbers are played to each ear. They then have to report these numbers back either by confirming the numbers heard in each ear (condition 1) or reporting the numbers in the order they were heard (condition 2). Broadbent’s findings were that generally participants found condition 1 easier than condition 2 and that participants generally couldn’t remember the ignored message in one ear when close attention was paid to the message in the other. However he found that with short messages, participants could access the neglected information a short time after, as it would be stored in the echoic memory. Physical features such as pitch and location were also found to be used by participants to track messages. From this, Broadbent concluded that switching attention back and forth between ears took up more cognitive resource and that the physical features were also used to filter information. He also proposed that all information is received in parallel and then sorted based on the perceptual characteristics. It is only then that the signal wanted is passed on for processing, anything else is blocked which protects again overloading of the semantic processor (Naish, 2010). Broadbent’s early filter theory has been criticised he argument being that if we haven’t extracted meaning out of the sensory information then how do we know to discard it as not relevant. Theories which contrast Broadbent’s early filter theory shall be looked at next.
Attenuation theorists suggest that only the shadowed message in dichotic listening tasks are fully processed, however everything else is not blocked, it is just attenuated. Treisman (1960) discovered that some participants, when shadowing a message, would say a word that had been presented in the unshadowed message. She found that this was usually because the word was plausible in the context of meaning in the shadowed message (Eysenck, Keane, 2010). This led Treisman (1964) to propose the attenuation theory which is that all information is perceptually processed in parallel then sent through for processing semantically. The attended message would be left untouched and then, in contrast to Broadbent, all of the other information would be attenuated and analysed, rather than blocked which explains why the attenuated message will sometimes be attended to (such as in the cocktail party effect where you can be attended to a conversation in one part of the room, hear your name in another part of the room and then switch your attention to that conversation) and also illustrates a flaw within Broadbent’s early Filter theory. Treisman used the dichotic listening task and shadowing to illustrate attenuation. Participants were asked to shadow a story being played in one ear ignoring what was being played in the other ear. She found that participants began to shadow the other ear if the story switched ears, contradicting Broadbent’s theory as they shouldn’t be aware of the story continuing in the other ear. Treisman proposed that this was due to priming where the temporary sensitisation of a word due to the presence of another causes the listener to assume what the next word will be. She also claimed that the location of the bottleneck was likely to be more flexible than Broadbent had suggested (Naish, 2010). Certain words, such as own name, may also be recognised easily in the unattended message. It was demonstrated by Underwood (1977) that the priming effect became stronger if there were several linked ideas used to prime rather than a single word. Underwood also concluded that sentences are processed as whole units but only when attended as the priming effect was also much stronger when the story starts in the attended ear and then switches to the unattended ear rather than vice versa (Driver, 2001). It has already been mentioned that Broadbent had found that information could be stored in the echoic memory and it is also true of the iconic memory when it come visual modality. He argued that if we could switch attention rapidly into the appropriate sensory buffer then it would be possible to process unattended information. Broadbent was, however, cynical about this as he believed that it took 500ms to shift attention, but it actually takes 50ms (Tsal, 1983 cited Eysenck, Keane, 2010). This means that shifting attention to that information in the echoic memory can be as efficient as shifting attention to the actual object. Treisman argues that the reason for this sporadic semantic processing of the unattended information is due to a leaky filter. In contrast, Broadbent argues that it depends on what is called ‘slippage’ by Lachter et.al (2004). That is attention shifted to supposedly unattended information, therefore it is not unattended (Eysenck, Keane, 2010).
There are three other studies that showed unconscious awareness of the unattended material. Corteen and Wood (1972) conducted an experiment where participants were conditioned with a mild shock to city names until they showed a Galvanic Skin Response (GSR). They still showed GSR to these words and cities they hadn’t been conditioned in, when they occurred in the non-shadowed message, although consciously they were not aware of them (Naish, 2010). Corteen and Dunn (1974) also conditioned their participants to certain words so that they would exhibit GSR upon hearing them. To demonstrate non-awareness, participants were instructed to make a conscious response when a conditioned word was played. They failed to do this most of the time when the conditioned word was played in the unattended ear and exhibited GSR. This suggests that some processing for meaning must take place and contradicts Broadbent’s early filter theory (Driver, 2001). Von Wright, Anderson and Stenman (1975) suggested that meaning can be processed without awareness. They established that a GSR was caused by related words, even synonyms, for the conditioned word by conducting a dichotic listening task where the participants were presented with two lists of words auditorially and asked to shadow one list and ignore the other. The findings were that this indeed happened and a GSR effect occurred when the words were in the non-shadowed list or were very similar sounding words (Eysenck, Keane, 2010).
As already mentioned, attenuation theory helps to explain the cocktail party effect and also Corteen and Dunn’s finding. It confirms that unattended information might be available for identification not meaning. There is also less perceptual information to be identified with special words such as own name. Attenuation theory also explains why participants will follow a message from one ear to the other as both messages are being processed however priming makes related words in attenuated messages easier to identify at switching point (Driver, 2001). It may also be possible to explain auditory attention (how we deduce distinct “sound objects” to attend to, such as one person’s voice among many) by extending visual attention theories that explain how we deduce and attend to objects in our spotlight of attention. This suggests the same neural mechanisms may lie beneath both types of perception (Shinn-Cunningham, 2008).
It is proposed by late selection theorists that all information is processed. It’s only the pertinent material that is made available Deutsch and Deutsch (1963) suggested that all messages are processed both perceptually and semantically, with the most important or relevant stimulus deciding the response. This puts the bottleneck much closer to the end of the processing system than both Broadbent’s and Treisman’s theories (Eysenck, Keane, 2010). Another theory suggested by Norman (1968) was that all information must be processed to the point where meaning is designated from memory (Naish, 2010). However, late selection theories were later falsified by developments in neuroscience. Neurophysiological studies took place which provided evidence against the theories of Deutsch and Deutsch. Coch, Sanders and Neville (2005), using the dichotic listening task, asked their participants to attend to one of two auditory messages. The task for the participants was to spot probe targets which were presented on both the attended and non-attended messages. Event-Related Potentials (ERP) were recorded from each participant. It was found that ERPs appeared 100ms after the probe was presented and was greater when the probe was presented on the attended message. This suggests that there was more assessing of attended probes rather than unattended probes. However, if processing happened in the way Deutsch and Deutsch theorise, then there wouldn’t be any difference in the ERPs recorded (Eysenck, Keane, 2010).
Each of the theories uses the dichotic listening task to provide evidence to support their theory. We know that attention is a limited resource and that we can’t focus on everything within our environment. The three major theories do give us an insight in to this. Broadbent’s early selection theory argues that the filter comes before the information is given meaning, but if this was the case then we wouldn’t be able to identify our own name when spoken in the unattended ear. Deutsch and Deutsch went on to argue that the filter happens after the information has been given meaning , so you register and assign everything a meaning then the filter decides what to pass on to conscious awareness. However, as attention is limited, this seems like a waste of resource to assign meaning to information that you will never need. So, Treisman’s attenuation theory fits nicely in between Broadbent’s early filter and Deutsch and Deutsch late selection filter and argues that the information into the unattended ear weakens but is not eliminated, that some of it will get through to the perceptual processes and be given meaning. The input from the unattended ear will still be given meaning but not as high priority as that in the attended ear. If, at this point, the unattended information is important (like hearing your names across a crowded room) then attention can be switched and you attenuate the previously attended to ear. The jury is still out on which of these theories is the ultimate when it comes to helping in our understanding of selective attention, however, each if the theories has been pivotal in how we view selective attention and this is important because attention is crucial to any other cognitive function that we perform.
Word count: 2046
Driver, J. (2001). ‘A selective review of selective attention research from the past century’,British Journal of Psychology,vol. 92, pp. 53–78.
Edgar, G. (2007). ‘Perception and attention’. In D. Miell, A. Phoenix, & K. Thomas (Eds), Mapping Psychology (2nd ed., pp. 3-50). Milton Keynes: The Open University.
Eysenck, M. & Keane, M. (2010). ‘Visual perception and attention’. In Cognitive psychology: a student’s handbook. (6th ed. Pp 153 – 201) Hove, Eng. New York: Psychology Press.
Naish. P., (2010). ‘Attention’. In Kaye, H. (Eds).Cognitive psychology (pp 29 – 62). Milton Keynes: The Open University
Shinn-Cunningham, B.G. (2008). ‘Object-based auditory and visual attention’,Trends in Cognitive Sciences,vol. 12, pp. 182–6.