PSGY1010 Cognitive Psychology 1 - Attention 2 Lecture Notes PDF

PSGY1010 Cognitive Psychology 1 Attention II: Early and Late Selection Dr Chung Kai Li [email protected]  Previously ▪ Early vs. late selection theories ▪ Early selection theories assume that unattended stimuli are not identified ▪ Late selection assume that unattended...

PSGY1010 Cognitive Psychology 1 Attention II: Early and Late Selection Dr Chung Kai Li [email protected]  Previously ▪ Early vs. late selection theories ▪ Early selection theories assume that unattended stimuli are not identified ▪ Late selection assume that unattended stimuli are identified, and meaning is analysed. ▪ So far, the evidence favours early selection ▪ But sometimes stimuli from the irrelevant channel are still identified 2 Reminder: irrelevant vs. relevant channel 3  Today’s lecture Learning objectives: ▪ Compare leakage, slippage and spillover ▪ Describe reasons for renewed interest in early selection ▪ Begin to understand how experiments are designed in psychology ▪ Explain effects of perceptual load in load theory of attention ▪ Evaluate the differences between Lachter’s early selection theory and load theory 4  Leakage Treisman (1960, 1964) ▪ Filter does not block information from the irrelevant channel, but it does attenuate it. ▪ Information from irrelevant channel “leaks” through the filter ▪ Attenuated information can activate concepts in long-term memory which leads to identification of stimuli 5  Slippage and spillover – demonstration water = attentional resources small container = relevant channel large container = irrelevant channel ▪ Aim: to pour water from the cup into the small container (or the relevant channel) 6 Slippage – demonstration 7  Slippage ▪ Metaphor: You cannot aim water (attentional resources) precisely enough (at least not all the time) ▪ Therefore, it is not possible to focus on the relevant channel all the time ▪ Consequence: if attention is not properly focused, then attention will slip to the irrelevant channel 8 Spillover – demonstration 9  Spillover ▪ Metaphor: You cannot stop pouring water until used up ▪ Similarly, we might argue that we cannot stop deploying attention until resource depleted ▪ Consequence: if the relevant channel needs less attention than that available, attention will “spill over” to the irrelevant channel 10  Slippage Lachter et al. (2004) ▪ For over 40 years researchers have assumed that there is identification without attention ▪ However, we think they are wrong! ▪ There is a need to ▪ Reinterpret old experiments ▪ Conduct new experiments ▪ We expect to conclude that Broadbent was in fact correct, there is no identification without attention 11  Reminder: Early Selection ▪ Early selection theories assume that there is no identification without attention ▪ Meaning is only analysed with attention ▪ Therefore, if stimuli in irrelevant channel has been identified then it has been attended 12  Reinterpreting evidence from last lecture ▪ Participants report words from the irrelevant ear when message switches to that ear ▪ Evidence that there is an own-name effect ▪ Evidence of “channel switching” ▪ Studies observed skin conductance changes when words associated with electric shocks were presented in unattended channel ▪ Evidence of subconscious processing ▪ All these findings seem to contradict Early Selection models ▪ However, Lachter et al. argued that none of these experiments had controlled for slippage, or involuntary attention to the irrelevant channel 13  Irrelevant vs. relevant channel Irrelevant Relevant ▪ The irrelevant channel is only channel channel irrelevant based on instruction Instruction: Instruction: Ignore this ear Attend this ear ▪ It has been assumed that this channel is not attended ▪ What if people sometimes attend to the irrelevant channel? 14  Own-name effect and working memory capacity Conway et al. (2001) ▪ Conway tested a group with high working memory capacity (WMC) and a group with low WMC ▪ In Group A, 20% noticed their own name ▪ In Group B, 65% noticed their own name 15  Own-name effect and working memory capacity ▪ The own name-effect depended on working-memory capacity ▪ Participants with low working memory: ▪ Noticed their name most frequently (Conway et al., 2001) ▪ Experienced difficulties focusing their attention (e.g., Colflesh & Conway, 2007) ▪ These results suggest that low WMC subjects are more likely to let their attention slip to the irrelevant channel ▪ Participants with high WMC are better able to control their attention ▪ So: the own-name effect is real; but this is an attentional problem 16  Channel switching ▪ Revision of the Channel Switching Experiment ▪ Message to shadowed ear: In the picnic basket she had peanut butter, book, leaf, root, sample, always ▪ Message to unshadowed ear: cat, large, day, apple, friend, every, select, sandwiches and chocolate brownies ▪ Participants will repeat information from the unshadowed ear ▪ Perhaps this occurs when subjects get confused when a semantically coherent message suddenly became incoherent ▪ To resolve the ambiguity, maybe participants reallocate their attention. ▪ This would be a form of slippage 17  Replication Studies ▪ In a replication study, the researcher tries to show the same thing again (sometimes with minor changes) ▪ Very important that we replicate findings, and is often not done frequently enough ▪ Replications are often unsuccessful ▪ Read: Nosek et al. (2015), Estimating the reproducibility of psychological science for more information. 18  Replicating electric shock conditioning study Dawson & Schell (1982) ▪ Attempted to replicate previous results that have shown changes in skin conductance as a result of words presented in the unattended channel ▪ They did find skin conductance changes, but in subjects who sometimes attended to irrelevant channel ▪ Failed to shadow relevant channel ▪ Recalled material from irrelevant channel ▪ This suggests that there might be slippage of attentional resources 19  Lachter et al. ▪ Hypothesis: “If we can exclude slippage, there will be no identification without attention.” ▪ How did they exclude slippage? ▪ Used visual stimuli, and presented relevant and irrelevant stimuli in different locations ▪ Presented the stimuli in irrelevant locations very briefly (55ms) ▪ How does this exclude slippage? ▪ Shifts of attention need time! It is not possible to shift attention to an irrelevant location in 55ms 20  The task: repetition priming ▪ In priming experiments, an irrelevant prime word (lower case) is presented shortly before the target word (upper case) ▪ Participants make a button press to indicate if the word is an actual word or a pseudo-word ▪ If a prime word is the same as the target word, this can speed up responses to the target word 21 How does repetition priming work? ▪ Seeing the prime activates the concept in memory ▪ It then becomes easier to indicate that the target “CAT” is a word 22  Repetition priming in different locations ▪ In classic repetition priming the prime and target are presented in the same location on the screen ▪ In Lachter et al the prime was presented in a different location to the target ▪ The target was always presented in the centre ▪ The prime was presented in the top of the screen ▪ Prime and target were presented for a very short duration (55ms) 23  Lachter et al. – results for same location ▪ Overall four conditions: 2x2 design ▪ Location: same as target or above ▪ Word identity: same (cat → CAT) or different (tip → CAT) ▪ Where the prime was in the same location then reaction times were ▪ Faster (smaller) when the prime and target were the same ▪ Slower (larger) when the prime and target were different 24  Lachter et al. ▪ Possible outcome 1 ▪ Same prime in different location does not affect RTs ▪ Possible outcome 2 ▪ The same prime in different location does affect RTs 25  Lachter et al. ▪ The actual result ▪ Outcome 1: no identification without attention ▪ Broadbent was right! 26  Was Broadbent correct…? Kouider et al. (2014) ▪ Training with auditory stimuli: ▪ left-hand response if word is an animal ▪ Right-hand response if word is a man-made object ▪ During sleep they presented new (untrained) words ▪ Result: participants prepared a response in their sleep! (shown with electroencephalography, EEG) ▪ They must have understood the meaning of the words! ▪ Tentative conclusion: Maybe there is identification without attention (but only in the auditory domain). 27  Spillover Lavie (1995) Combined assumptions from late and early selection. ▪ Perceptual processing is automatic ▪ Typical assumption from late selection theory ▪ However in contrast to late selection theory she argued that perceptual processing is capacity limited ▪ Typical assumption from early selection theory ▪ This theory is sometimes portrayed as a “hybrid” theory which combines assumptions from early and late selection theories ▪ But really an early selection theory because it assumes that there is no identification without attention (Benoni & Tsal, 2013) 28  Lavie (1995) – spillover ▪ Assumption: identification of irrelevant stimuli only if processing of relevant stimuli does not exhaust available resources Total Needed Available resources relevant irrelevant 29  Lavie (1995) – capacity limitations ▪ If processing of relevant stimuli consumes all available resources then there is no identification of irrelevant stimuli Total Needed Available resources relevant irrelevant 30  Lavie & Cox (1997) ▪ Task: detect target letter in circle (always X or N), press left for N, press right for X ▪ The flanking distractor letter presented at the side of the screen (also X or N) is irrelevant ▪ Compatible trials: the target and flanker are identical ▪ Incompatible trials: the target and flanker are different (as in the example) 31  Lavie & Cox (1997) – manipulating perceptual load Two conditions ▪ Low perceptual load ▪ Target immediately visible, no search required ▪ Attentional resources “spill over” to flanking letter (the irrelevant channel) ▪ High perceptual load ▪ Target hidden among neutral distractors (not associated with a response), search for target required ▪ Central circle (the relevant channel) consumes all attentional resources ▪ No spillover to irrelevant channel 32  Lavie & Cox (1997) ▪ Hypothesis: compatibility effect for low perceptual load trials, but not for high perceptual load trials ▪ What is a “compatibility effect”? ▪ Difference in reaction times (or error rates) between incompatible trials (e.g., target X and flanker N) and compatible trials (e.g., target X and flanker X) 33  Lavie & Cox (1997) ▪ Low perceptual load condition ▪ There was a large compatibility effect ▪ Average of 40ms difference ▪ Suggests that the flanker in the irrelevant channel has been identified ▪ High perceptual load condition ▪ There was a small compatibility effect ▪ Average of 4ms difference ▪ Suggests that flanker in the irrelevant channel has not been identified 34  Lachter vs. Lavie ▪ Both studies suggest that there is no identification without attention, therefore they are early selection theories. ▪ Lachter: if attention properly focussed, no slippage → Processing of irrelevant channel avoidable ▪ Lavie: capacity of perceptual attention limited, but not under voluntary control → Low perceptual load: processing of irrelevant channel unavoidable 35 Thank you Any questions?