Metacognition Flashcards
What is metacognition?
- Processes involved in monitoring and controlling performance on a task
- an awareness of one’s own thought processes and an understanding of the patterns behind them.
- How a human adapts memory performance using the things they know about their memory
Why study metacognition?
- Traditional experimental memory paradigms don’t capture this aspect of cognition
- Early studies looked at the amount they could recall not the way in which they did it
- Traditional memory paradigms don’t reflect the operation of memory in real world environments
- What we learn could be used to improve many decisions based on information accessed from memory
what are the 2 key assumptions the meta-cognitive approach is based on
- humans have the ability to monitor their own cognitive experience
- This subjective experience that arises from monitoring plays a causal role in determining current and future cognitive operations
how does monitoring and control work (object level and meta-level)
examples
object level > meta level = monitoring- we monitor our subjective experience of feeling
e.g. check you understand what you are reading
meta level > object level = control- determines current and future cognitive operations (how long you might continue thinking about it for
e.g. re-reading a paragraph
what is monitoring, how is it measured
- Processes that allow us to observe and reflect on our cognition
- Monitoring means the ability to judge successfully one’s own cognitive processes, and control means the ability to use those judgments to alter behaviour
- Measured by asking people to report their own monitoring
what are 4 measures of monitoring?
- ease of learning
- judgment of knowing
- feeling of knowing
- confidence
ease of learning?
before you learn each item, how difficult will it be for you to learn it
judgment of learning?
for each of the items you have learned, how likely is it you will be able to recall them later
feeling of knowing
tip of tongue state
For items that you are not able to recall, how likely is it that you would recognise the item from among a set of possible items?
confidence?
• For items that you recall/recognise, how likely is it that your decision is correct?
what is control
- Conscious and non-conscious decisions we make based on the output of our monitoring processes
- Revealed by behaviours (measurement of what people do)
name measures of control
- Self-paced study time (measure how long someone looks at stimulus for)
- Response time (recognition and recall)
- Quantity of information reported (things they say)
- Veridicality/ correspondence of information reported e.g. an item could say a person wears a red jumper, how does this correspond with original item shown
- Grain-size of information reported (how detailed a report is, we can ask people to bet on how wrong/right they think they are)
- Many of them are verbal
Example of metacognition in action: studying for exams
• You have to decide when you have learned enough
• You could measure:
> Monitoring: judgements of learning
> Control: amount of time spent studying each item
what did hart (1965) look at
looked at feeling of knowing judgements predict performance on recognition test
what did Underwood (1966) look at
looked at whether ease of learning judgements predicted recall performance
what paradigm Hart’s (1965) study use
what type of questions were did participants answer
• Recall-judgement-recognition paradigm was used
• The focus on general knowledge questions (semantic memory) NOT episodic
• Participants answered 50 or 75 general knowledge questions e.g. what is the capital of Japan?
• Participant fails to recall the answer- how likely will you be able to recognise the answer if we present you with a set of possible answers?
> Either gave Yes/no or ratings on a scale of 1-6 (wouldn’t recognise, through to would recognise it)
> Multiple-choice recognition of all items e.g. a) Kyoto, b) Tokyo, etc
what were the findings from Hart’s (1965) study
what does this tell us about people’s metacognitive knowledge?
When they were correct, more likely to say they knew the answer (felt they knew)
> There is quite a lot of variability across participants
> FOK=items participants felt they would not know, percentages were lower
> M=.66 vs. .38
> Some people have good metacognitive knowledge, and some people don’t
what were the findings from Hart’s (1965) study
what does this tell us about people’s metacognitive knowledge?
For items they would know the items, they had higher correct recalls
For items they would not know or recognise them, they had fewer percentage of items they got correct
> There is quite a lot of variability across participants
> FOK=items participants felt they would not know, percentages were lower
> M=.66 vs. .38
> Some people have good metacognitive knowledge, and some people don’t
what did Underwood’s (1966)- ease of learning judgements (EOL) study comprise of
what paradigm was used
what did participants do
what were the 2 conditions
- Paradigm in this experiment involved 27 “triagram” (3 letters), easy -> difficult: BUG, CES, XFH
- Participants judged the speed with which they would learn each trigram
- Surprise recall test and Intentional learning conditions: over 6 study-recall blocks
what were the results of Underwood’s (1966)- ease of learning judgements (EOL) study?
- Key question: Do judgements predict recall performance? Answer is yes, not too badly
- Correlation between average ‘speed of learning’ judgements and accuracy
- Surprise recall r=.77
- Intentional recall r=.91 (when people knew they would have to recall the information)
- Some people had really strong correlation, some had poor correlation
- Reasonable correlations for individuals too (mean r of .48), but some individual differences
- Metacognitive judgements are of intermediate accuracy (above chance but far from perfect)
what is the direct access view, implication of this?
• Judgements are made on the basis of features of the target that can be accessed or retrieved
> E.g. people access some notion of memory strength for a learned item
> E.g. Hart (1965-67): for a non-recalled item, a memory monitoring process ascertains whether the item is stored -> FOK
Implication: weakly stored information should not be predicted to be more recallable than more strongly stored information
what is the inferential view, describe what inferences may be
• Judgements are based on a host of cues and clues generated from 2 sources:
1. Generated by act of learning or remembering
2. From knowledge-specific info about own memory
Inferences may be:
> Largely non-conscious, revealed as a “feeling” or
> Based on a theory about what makes information memorable
• E.g. ease of retrieval: amount of information retrieved
what is the hybrid view and what 2 cues does it consist of
Koriat said metacognitive judgements are based on experienced based cues and information based cues
what are experienced based cues
• Sheer feeling arising from some aspect of remembering, learning or failing to remember
> Interpretation of what that means
> E.g. looking at a photo of someone famous but cannot recall their name
> Tip of the tongue is an example of experienced based cue
what are information based cues
• Our knowledge of our own memory
• A- priori theories (knowledge that is acquired independently of any particular experience) and analytical inferences about the impact of various factors on memory e.g.:
> The longer a stimulus is present for, the more likely I will later remember it
> The longer the delay between learning and test, the worse I do on the test
> The more I practice an item, the more likely I’ll remember it
what does memory allow us to do
- navigate our way around places we’ve never been before
- communicate effectively because we’ve developed language based on memory
- memory not based on HOW MUCh we remember BUT on WHAT we remember and verticality of those memories
what is a list learning paradigm
> it looks at memory performance through measuring the percentage of items recalled/recognised
memory is measured in terms of its effectiveness by the percentage of items recalled and NOT by what is recalled
Koriat and Goldsmith highlight behind this list learning paradigm is a metaphor for what memory is
what is the storehouse metaphor
- memory is the storehouse of info that we can later access and retrieve
- A lot of the theories in this area use language such as a memory trace, memory vector or memory item
- Forgetting under this paradigm is thought of as a loss of elements, so we lose items from memory under this paradigm
what are 2 traditional approaches to memory
- list learning paradigm
2. storehouse metaphor
correspondence approach to memory: every day memory research/what is it exemplified in
correspondence between what is inside our head and what we ACTUALLY perceive in the world
- exemplified in eye-witness research
- show people a video, ask them to recall it to best of ability
- interested in correspondence of what they report and they originally saw
what is the correspondence metaphor (accuracy oriented)
- Memory is about the past event
- Focus is on the accuracy of the report
- Memory is assessed as an output bound manner rather than an input bound manner
- Forgetting is a loss of correspondence
how do we measure correspondence
You could calculate the accuracy by dividing the number of correct words by the total number of words that they were shown
OR
you could look at their accuracy by dividing the total number of correct words they produced by just the numbers that were produced so you’re not taking account of blanks left
which approach to measuring correspondence results in higher accuracy and why
total words correct and total words produced (doesn’t account for blank words) produces higher accuracy
as we’re not actually saying they were wrong when they failed to reproduce a couple of items (accuracy=8/12=66% vs. accuracy =8/10=80%)
this approach is more important for real world memory
compare how items would be measured under storehouse metaphor and correspondence metaphor
• If a person fails to remember the word “mill” and left it blank, they haven’t actually reproduced anything
- In the storehouse metaphor for memory you would say that its wrong
- In a correspondence, you would leave that blank and wouldn’t count that at all and just leave it out
• For the studied word “show”, the word “point” was reproduced
- In the storehouse metaphor and correspondence metaphor this is considered as wrong because it does not correspond with the original word
• For the studied word “plant”, nothing is reported
- For storehouse metaphor this is wrong
- For correspondence, it is left blank
• Hence, it is apparent there is a difference in the way that we’re counting under different metaphors
what is the recall-recognition paradox?
when you get two different types of results depending on the metaphor for memory that you use
Namely, the people were forced to produce items from memory in the lab-based studies but they were given free report option in the eye witness or more real world studies
> people were able to not report things and could leave items blank if they wanted to
> Finally, there was a difference in the format that was used
> They argued that there was not really a paradox and that it just came down to methodology and the way things were measured
Why does the recall recognition paradigm occur?
3 things
Memory property under consideration
> accuracy versus quantity
Report option
> forced versus free
Test format
> recall versus recognition
what did koriat and goldsmith find in their study
