Please enable JavaScript.
Coggle requires JavaScript to display documents.
Attention (Y1) - Coggle Diagram
Attention (Y1)
Models of attention
Early filter model - Broadbent, 1958 -
- First detailed information processing model
- Low-level sensory information from attended signal (e.g. attended right ear) and unattended signal (e.g. unattended left ear) gets into sensory store, but only specific information is selected for processing
- Selection based on physical properies (location, colour, etc)
- Various specialised channels perceive information from the environment bringing information to the buffer store and it is then passed through a filter which selects based on physical properties - chosen stimuli processed, rest are disregarded
- Inputs (attended and unattended message) -> Sensory store -> selective filter (based on physical properties such as pitch, or unattended signal blocked) -> Bottleneck -> Higher level processing -> working memory
Issues with the early filter model -
- Unattended yet highly familiar information can slip past the filer e.g. one's own name
- Attention not allocated in all or none fashion
- No ecological validity
Attenuation / attenuator model - Treisman, 1964 - varying weight of processing to different pieces of information
- Selection still based largely on physical properties, and to a smaller extent familiarity or relevance
- Different mechanism for filtering stimuli at the input channel - process the different stimuli to different degrees
- Attenuators are also influenced by previously analysed material - possible some stimuli, in particular highly familiar or semantically relevant stimuli in context reach the semantic level although they are not related to the initially attended cluster of stimuli
- Attended information gets to be processed semantically
- Inputs are the same as early filter, but after the sensory store the attenuating filter based on physical properties takes attenuated message down the bottleneck to a hierarchy of analysers (syllables, words, grammar and semantics) -> working memory
Issues with early selection models -
- Even unattended messages can be processed to the level of meaning
- Both models specify that selection is early at the sensory memory level, but the attenuator model allows for graduation of filtering that the filter model does not - explains how we fear familiar stimuli and context-relevant stimuli
- Both see unattended stimuli as left to decay in short term memory - does not influence behaviour
Late selection models - Deutsch and Deutsch (1963) -
- suggested that selection happens at the level of memory - stimuli are processed for meaning and assigned corresponding representations based on previously stored knowledge
- require all information and stimuli to be recognised in order for a suitable action to be chosen
- Unattended stimuli have an unconscious effect on behaviour
- Information must be processed rather deeply first - wasteful of cognitive resources
- Stimulus -> sensory registration and storage -> perceptual analysis -> response selection
Subliminal priming - cross judgement task (Mack and Rock, 1998) - judge whether the horizontal or vertical stroke of a cross presented in one of the quadrants is longer
- Participants dd not detect the briefly presented word at another quadrant
- Able to choose out o 5 alternatives better than chance
- However, little evidence that subliminal priming - not aware of the stimuli that prepared us for the upcoming one
- Priming - preceding stimulus makes processing of the next stimulus easier
- Unfavourable to early selection models in the same way dichotic listening is unfavourable - does not allow stringent control of the stimuli the participants attends at the sensory level
Summary - models of selective attention differs on when selection occurs (early or late), whether allocation is all or none and whether the unattended information is processed at the level of physical property only or meaning also
- All models are supported but have issues
- Perhaps attention is more flexible and situation dependent - could use a range of models
Cherry (1953) - dichotic listening - different auditory message is presented to each ear at the same time, and when the voices were the same (same physical properties) people had difficulty separating the messages using meaning only, which suggests they are initially processed
- interested in how much is retained from the unattended message
- Had to shadow the target message (repeat it aloud) - most physical properties of the unattended message were noticed, such as reverse speech, language change and sex change
-
-
-