Moray (1959) - Auditory Attention

• cocktail party effect

  → cherry (1953) described the 'cocktail party effect' which describes our selective attention - we're good at tuning into 1 voice // conversation over another ,, tuning rest out

  → however some information can break through attentional barrier ,, E.G. hearing your own name ,, or a mother who hears her child crying
• AIM

  → whilst we pay attention to 1 message ,, we set up a 'block' on any other message ,, only focusing our attention on selected message

  → Moray was interested in what types of message could penetrate this block & be paid attention to by participants

  → so Moray aimed to find out which factors could enable an unattended ,, dichotically presented message to be noticed
• dichotic listening tasks

  → in these tasks ,, 2 different auditory stimuli or 'messages' ,, are presented into different ears through headphones

  → participants have to 'shadow' // repeat one of these messages aloud as they listen

  • SHADOW MESSAGE ; attended message (message repeated verbally) - READ ALLOWED!
  • REJECTED MESSAGE ; unattended // blocked message
• method - 3 experiments

  → in all of experiments ,, participants complete a dichotic listening task where they had to shadow 1 message ,, which were played a tape in same male voice at rate of 130-150 words per minute

  → volume of shadowed & rejected messages was kept the same

  → before each experiment ,, participants completed 4 trial-shadowing tasks so that they could practice shadowing before study
• sample

  EXPERIMENT 1 → Moray didn't make note of amount of participants in this

  EXPERIMENT 2 → 12 participants

  EXPERIMENT 3 → 2 groups of 14 students

  • ALL PARTICIPANTS WERE UNIVERSITY STUDENTS or RESEARCH WORKERS OF BOX SEXES
• experiment 1

  → shadow message (prose message)

  → rejected // blocked message (word list)

  • RECOGNITIN TASK → 30 seconds after completing shadow task ,, participants complete recognition task of 21 words (7 from shadow ,, 7 from rejected list ,, other 7 controls)
  • IV → message type (shadow // rejected message)
  • DV → number of words recognised from each message out of 7
• experiment 1 - results


  → participants were much better at recognising words from shadowed message than from rejected message

  → participants remembered very few of the words from rejected message ,, despite having hard it 35 times

  
• experiment 2

  → PROCEDURE - participants listen to 10 passages ,, recieved instructions within all of them. for some passages the instruction was preceded by participants name (affective) ,, for others it wasn't (non affective)

  → fictional passage - right (shadowed message - right first)

  → different fictional passage - left

  • IV → whether instructions in passage were preceded by participant's own name or not (affective // non affective instructions)
  • DV → whether or not they heard instruction (so swapped ears)
• experiment 3


  → RIGHT & END - message (with digits added towards the end)

  • IV → instructions given to participants

  → either given general instructions that they would be asked questions about shadowed message // specifically told to remember as many digitsas possible

  • DV → number of digits recalled
• experiment 2 - results ~ 2


  → inferentual statistical test showed that participants heard affective instructions significantly more than non-affective instructions at level of p<0.01

  • this means that there was less than 1% probability the difference was due to chance factors therefore ,, there's a 99% probability that presence of their name did cause ppts to hear instruction in rejected message

  
• experiment 2 - results ~ 1

  → participants did 10 different shadowing tasks ,, & part way through the shadowing task they would receive an instruction in rejected message

  → instructions would either be affective (i.e. preceded by participants own name) // or non-affective (i.e. not preceded by their name)

  → participants rarely heard the instructions in rejected message when they weren't preceded by their name (heard 11% of the time ,, non-affective instructions)

  → however ,, they heard instructions half (51%) of time when they were preceded by their name (i.e. affective instructions)

  
• experiment 2 - results ~ 3

  → moray noted that ppts were given instructions at start of passage to expect instructions to change ears ,, i.e. they were primed ,, there was slight increase in how often ppts heard instructions in rejected message

  → moray conducted the third experiment to test this idea
• experiment 3 - results

  → there was no significant difference between number of digits correctly recalled in each condition

  → shows that participants can't be primed to respond to digits in rejected message ,, unlike we do spontaneously with our name (E2) when participants had been specifically told they would hear digits ,, this wasn't 'important' enough to break through attentional block

  
• conclusions ~ 1

  → when a person directs his attention to a message from 1 ear & rejects a message from other ear ,, almost all of verbal content from rejected message's blocked

  • EVIDENCE - comes from experiment one → when participants recalled 70% of words in message they were paying attention to ,, but only 27% of words from rejected message
• conclusions ~ 2

  → rejection's apparent even when message is repeated many times

  • EVIDENCE - comes from experiment 1 → when word list was repeated 35 times as rejected message ,, yet despite, this only 27% of words were recalled from this message
• conclusions ~ 3

  → subjectively important messages ,, such as person's own name can penetrate the block. so we may hear instructions in rejected message if instructions contain our own name

  • EVIDENCE - comes from experiment 2 → where participants heard instructions containing their own name were heard in rejected message 51% of time ,, but only 11% of time without their name
• conclusions ~ 4

  → it's very difficult (though perhaps not impossible) to make neutral material important enough to penetrate the block

  • EVIDENCE - comes from experiment 3 → where participants struggled to remember digits from rejected message even when primed to listen out for them
  • HOWEVER ,, it's not completely impossible as in EXPERIMENT 2 ,, participants still heard neutral // non-affective instructions 11% of time
• evaluation ; strength

  → conducted a highly controlled lab experiments where he controlled extraneous variables such as volume of messages - were also checked to ensure they were spoken at a constant speed - control leads to higher validity

  → highly strandardised procedure means experiments replicable so results can be checked for reliability using test-retest method

  → collection of quantitative data

  • in E1 ,, number of words that participants could recall from each message allows comparisons to be made across groups ,, I.E. whether more words were recalled from shadowed or rejected message
• evaluation ; weakness ~ 1

  → experiment lacks ecological validity - rarely in real life would a person need to listen to a message and repeat it aloud. however ,, dichotic listening tasks somewhat reflect real life ,, E.G. when trying to follow once source of information such as a conversation ,, whilst ignoring background noise or chatter. also use of everyday materials ,, such as using light fiction ,, & use of the individuals own name help to raise ecological validity
• evaluation ; weakness ~ 2

  → as participants were aware they were taking part in an experiment ,, there's the possibility that they responded to demand characteristics & tried to influence the outcome of study.

  • they may have thought they were '' supposed '' to not remember anything from rejected message & so reported not recalling content from these messages when in fact they did - would lower validity of research
• evaluation ; weakness ~ 3

  → sample was made up of students & research workers ,, which as a group may have a higher level of cognitive ability than rest of population. therefore ,, they may outperform general population on tasks that require cognitive skills ,, such as selective attention tasks. along with the small samples used in this study ,, this means that sample may not well represent the target population & therefore the results cannot be readily generalised beyond the sample