Attention: early and late selection

WIlliam James (1890) saw selectivity as the hallmark of attention. Selective attention is still the focus of most modern research on attention

Selective attention - refers to the ability to focus attention on a specific stimulus, such as an object, location, or message

Limits to what we can perceive

Mack and Rock (1998) carried out research on:

• inattentional blindness: failure to notice an unexpected object that is clearly visible
• change blindness: failure to notice a change in a visible object
• These failures are not to do with vision deficits, but rather limits in perceptual processing capacity and lack of atention

Why do we need selective attention?

• Unless we pay attention, we are not aware of what we see of hear
• We sometimes overlook our friends in a crowded room, and sometimes fail to notice when a friend gets a haircut
• Selection must take place to focus our limited capacity cognitive processing resources on a subset of information and ignore the rest

Early selection - selection takes place at early stage of perception with limited processing of ignored inputs

Late selection - selection takes place at late stage of perception with extensive processing of ignored inputs

Broadbent's (1958) FIlter theory of selective attention

• Broadbent's interest was fuelled in part by his knowledge of the problems faced by radar operators in WWII
• They would be required to listen to and communicate with several different pilots whose voices were being relayed over a single loudspeaker all at once
• Broadbent proposed that attention acts as an early selective filter to prevent information at a late stage of processing
• This filtering occurs by selecting relevant information on the basis of simple physical properties during an early stage of processing

Support for Broadbent's model

Broadbent (1958) investigated the extent to which we process unattended information under experimental  conditions using 'selective shadowing'

• In the unattended ear - participants failed to notice a complete change in the language spoken
• Participants did notice switch from male to female or switch from speech to 400 Hz tone 
• Spering (1960) showed that participants typically report 4/5 items from brief visual arrays containing 12 items - a clear demonstration of limited capacity processing
• Von Wright (1970) showed that physical cues were more effective in allowing selective report than semantic cues - thus attentional selection operates on basic physical features - evidence for early selection

Problems for Broadbent's model 

• 'shadowing' studies often used retrospective questioning e.g. surprise questions about a non-shadowed message at the end of shadowing

Deutsch and Deutsch (1963) Late Selection Theory

• Parallel processing of all inputs to the point of full semantic analysis
•  Selection is based on the importance or relevance of stimuli
• The bottleneck in processing is placed after semantic analysis but before entry into awareness and memory-based systems for the control of behaviour

Support for Deutsch and Deutsch

Mckay (1973)

-Participants favoured the interpretation of the sentence suggested by the unattended/unshadowed word

Von Wright, Anderson and Stenman (1975)

• Certain words were paired with electric shocks
• Because of conditioning , the words then caused a fear response presented on their own, measurable by galvanic skin response or GSR.
• using the selective shadowing task, Von Wright and colleagues found that conditione words and synonyms of these words in the ignored ear still produced GSR reponses, even without awareness
• They concluded that some meaning must be processed from the non-shadowed message and that Broadbent's model cannot be entirely correct

Tipper (1985)

• Negative priming effect - participants show slower reaction times for recognising an object if that object has been previously ignored
• Unattended visual objects may undergo full perceptual processing and object recognition but then be actively inhibited

Early selectionists considered the evidence used to support 'late selection' as reflecting failures of attention

Revised Early Selection Model 

• Treisman argued that instead of a strict filter, processing was weakened for the unattended stimuli
• Because it is weaker, we can usually ignore it. However, this weaker input may sometimes trigger awareness of high level meaning in cases where the high level meaning is 'primed' - e.g. information that is consistent with current expectations or that is personally relevant

Support for Treisman

• Moray (1959) showed that people would notice their names if they were presented in the unattended ear during a shadowing task, even if it had the same physical characteristics of the other unattended side
• other studies suggested that 'rude and nasty' words were also noticed
•  we notice these words because we've processed their meaning

Weakness - It is underspecified in that it does not precisely detail the attenuation process

 

Summary of models

Broadbent FIlter theory

- Selection is based on simple physical characteristics. WHat is not selected is turned out almost completely

- Only superficial features of unattended messages are reportable

Deutsch and Deutsch Late selection theory

- No capacity limitations in perception; all information is processed completely - bottleneck then prevents some of this information from reaching awareness

Treisman Attenuation theory

- selection is based on simple physical characteristics. The filter attenuates rather than blocks information from the unattended message