Aim
to test Cherry’s dichotic listening findings in relation to:
- amount of info recognised in unattended msg
- effect of hearing your name in unattended msg
- effect of instructions to identify a specific target in unattended msg
Research method
Lab experiment
+ : highly controlled environment, eliminates extraneous variables, ez to replicate so reliable
- : artificial, low ecological validiity, difficult to set up lab procedures where ppl would behave as they would irl, demand characteristics
1) Repeated measures
2) Independent measures
3) Independent measures
Participants
- m & f undergraduates + research workers
- exp 1: not given ppts
- exp 2: 12 ppts
- exp 3: 2 groups of 14
Controls
- same recorder speed 130 wpm
- same volume on either side, same tape recorder
- ppts given 4 passages to prose shadow for practice
Procedure 1
1) Simple words short list repeated in 1 ear while ppts shadowed (repeated) a msg in other ear
2) list repeated 35 times, faded in after shadowing began + out as msg ended
3) ppts asked to report all content they could from rejected msg
4) Given recognition test, included: words from shadowed, unattended & words in neither
5) Gap between shadowing end + recognition = 30s
P1 Variables
- IV: Shadowed, unattended msg
- DV: words recognised correctly in rejected msg
- CV: Recognition test
Procedure 2
1) Ppts shadowed 10 short passages, asked to make few mistakes as possible
2) Instructions inserted in non-attended ear at start, within passage or both eg. listen to right ear
3) 6 instructions given within passage eg. change to other ear. 1/2 prefixed by name (affective instruction)
4) 4 passages; ✕ instructions given within msg
P2 Variables
- AIM: find limits of attentional blocks efficiency
- IV: whether instructions were prefixed by ppts name (affective) or not (non-affective)
- DV: no. affective instructions perceived
Procedure 3
1) 2 groups of 14 ppts shadowed 2 of 2 simultaneous dichotomy msg
2) In some msgs, digits inserted towards end of msg. Sometimes in both msgs & sometimes in 1. Positions of numbers in msg and controls w no numbers were randomly inserted.
3) 1 group told they’d be asked Qs ab shadowed msg content at the end of each msg & others instructed to memorise all numbers
P3 Variables
- AIM: to show instructions may alter chances of material in rejected msg being perceived
- IV: where ppts were told b4 that they had to answer Qs ab shadowed msg at end of passage or memorise numbers
- DV: no. Digits correctly reported
Results 1
words presented —> mean no. Words recognised /7
- shadowed —> 4.9
- rejected —> 1.9
- for 1st time in recognition test —> 2.6
Results 2
- presence ✕ name can cause instruction to be heard ; affective content in unattended msg can break thru attentional barrier
instructions followed
> 20/39 w name
> 4/36 ✕ name
Results 3
- diff between mean no. Digits reported under 2 conditions (told they’d be asked ab content or to listen to digits) was ✕ statistically significant
- ; even when alerted to possibility of hearing digits, perception ✕ affected
Conclusions
- if ppts direct attention to 1 ear + reject other, almost none of verbal content of rejected is able to penetrate attentional barrier
- short list of simple words presented as rejected msg shows no trace of being recalled even when presented many times
- subjectively important msgs eg. Name can penetrate barrier ~ person will hear instructions containing own name in rejected
- v difficult to make ‘neutral’ material important enough to break thru barrier set up in dichotic shadowing
Type of data
Quantitative
1) no. Words recognised correctly
2) no. Affective instructions perceived
3) no. Digits correctly reported
✓ can draw table
✓ easy to analyse/compare
✕ lacking human insight, why?
Validity
- high internal ; controls (eg. same vol) ensured only IV successfully impacted DV ~ tested what it was set out to test
- low external ; not realistic but fiction paras could increase realism
Reliability
- high, replicable
- extraneous variables limited by controls (eg. sound speed 130wpm - words per min) ~ standardized procedure
Sampling bias
Ethics
- ✕ generalise ; not everyone undergraduates/researchers, used to being tested + need certain intelligence lvls
✓ m + f but small sample ~ ✕ represent all society
no ethical issues
Ethnocentrism
- ethnocentric ; all from same area (single Western uni)
- differences found between languages ; monolingual speakers had right ear advantage over bilingual
- bilingual ppl struggle w noisy environments
- BUT Swedish/Finnish perform better
Practical Applications
- helps w teaching methods in skl
- improving attention
- testing pre-verbal children’s hearing ~ understanding hearing impairments
- testing coma patients
