Lecture 2 Attention Flashcards
What did cherry find with binaural presentation of messages?
This is when both ears receive both messages. Same voice. Diff content.
Needed to repeat the messages 10-20 times until task could be completed
Shows us that cocktail party problem is a dichotic listening thing, as we can use enviro cues such as spatial cues to localise sounds
Also sound waves arrive at left and right ear at diff times - phase differences) so we can kind of know where sounds are relative to each other
Binaural task took all this away
Criticisms of cherry?
Only looked at what’s remembered… Should be interested in what’s perceived
Confounds perception and memory (can’t distinguish between the two)
Maybe perceived by forgotten?
What is and who came up with filter theory?
Broadbent, 1958
- that attention acts as a selective filter which stops the brain from being overloaded by all the stimuli in the environment
- there is a limited capacity channel which are cognitive processes which turn speech sounds into an understanding of speech
FILTER IS BEFORE LIMITED CAPACITY CHANNEL, protects it from overload.
Stm temporarily stores it all
What is the evidence for filter theory.
Digit span experiment
6 numbers:
People were bad at temporal order (the order the numbers were heard in both ears) - need 5 filter switches - memory may decay by the time they switch
People preferred ear by ear basis - only uses 1 filter switch
What study shows the flaws of filter theory?
Dear aunt Jane experiment - Gary and wedderburn, 1960
Split span experiment
Stimulus was 3 word meaningful phrases
People preferred to recall semantic content, not presentation ear
Showed that meaning is extracted before the filter?
What’d moray (1959) find?
Supported early detection theory
Did the dichotic listening and shadowing task, but embedded the subjects name in the unattended channel - people detected it.
What’s tries mans attenuation model
He argued that the unattended channel gets through to the LTM in an attenuated way…
Only attended stuff gets fully activated so were aware
It’s like turning down the volume
Biased by content, message salience. Highly salient or semantically related material gets through filter and shifts attention
What did triesman and geffen find for early selection?
Tried to test what people perceive before the forgot it
Had people make a tap when they heard taps in either channel..
They were better on attended channel, and detected less on ignored channel —- supports attenuate model!
What did treisman and Riley make adjustments to in the tap study for early selection theory?
They changed the “tap” word spoken in different voices in the two channels
People could now detect both channels equally due to being able to presttentively process physical features
What are criticisms of early selection
- filter must be complex to be able to respond to semantic content… And distinguish from unrelated stuff
- so maybe makes sense that filter is after the semantic system
Evidence for late selection?
Semantic processing on unattended channel
McKay
Shadowing experiment followed by Re-cognition experiment
Ignored passage had words relate to ambiguous word in recognition stimuli.. Ignored passage affected their understanding..
Von wright, Anderson and stenman - 1975
Conditioned galvanic skin response with word group
Generalised to ore the words in the category
GSR PRODUCED IN ignore channel
Norman 1968 late selection….
What’d he say
Late selection can be bottom up from stimuli to us
Or top down - intentional activation - leads to variable called pertinence
What did cherry find with dichotic listening and shadowing tasks?
Dichotic listening and shadowing.
Two passages of continuous speech on each ear, asked to attend to one and ignore the other.
Person instructed to repeat message 1 as the heard it to ignore the second one.
People had no memory of the unattended message, and didn’t notice when it was switched to German, reversed.. But noticed if the voice changed.
So found that only physical features of unattended message perceived, semantic content not perfected.n
