7 – Memory 2

Question 1

Describe the traditional multistore model of memory (Atkinson and Shifrin, 1968)?

Answer 1

1. Sensory stores, e.g. iconic memory
2. Short-term memory
3. Long-term memory

Question 2

In the traditional multistore model of memory (Atkinson and Shifrin, 1968), how does information get from the sensory stores to short-term memory?

Answer 2

Through attention.

Question 3

In the traditional multistore model of memory (Atkinson and Shifrin, 1968), how does information get from short-term memory to long-term memory?

Answer 3

Through rehearsal.

Question 4

In STM letter-span tasks, how many items do people normally remember?

Answer 4

7 +- 2 is magic number.

Question 5

In STM letter-span tasks, what if they have to count out loud while remembering?

Answer 5

They remember fewer letters. They can’t rehearse by mentally repeating numbers, as counting uses the same code –speech system – as you use to rehearse.

Question 6

In STM letter-span tasks, what if letters sound the same?

Answer 6

It’s harder. Because speech-based phonological code used for rehearsal.

Question 7

In STM letter-span tasks, what if words are used instead of letters?

Answer 7

It’s easier. But long words harder than short –because of phonological rehearsal. Ditto for unconnected words vs connected, which allow for use of semantic cues.

Question 8

What is the capacity of STM according to the traditional multistore model of memory (Atkinson and Shifrin, 1968)?

Answer 8

7 +-2

Question 9

How quickly does information decay from STM?

Answer 9

Within 30 seconds if not rehearsed.

Question 10

In the traditional multistore model of memory (Atkinson and Shifrin, 1968), what is the capacity and rate of forgetting of the LTM?

Answer 10

Capacity – unlimited

Forgetting happens due to interference rather than decay

Question 11

According to the traditional multistore model of memory (Atkinson and Shifrin, 1968), how does information end up in LTM?

Answer 11

Information first must be attended to and registered in iconic memory. Then it must be rehearsed for it to be transferred to LTM.

Question 12

Give two examples why repetition is not sufficient for encoding?

Answer 12

1. Craik and Watkins (1973) gave list of words to Pps, told only to remember last word with the letter G. They could estimate how many times each G word is rehearsed based on the number of items between each G word. Found no difference in level of recall for words as function of how many times they were rehearsed.
2. Kellas et al. (1975) Memory better when people rehearse silently rather than out loud. It was hypothesised that when people didn’t have to engage in overt rehearsal, they were also doing OTHER things to these items –e.g. semantically processing and integrating –that accounted for better performance.

Question 13

What are the three levels of encoding in Craik and Lockhart’s levels of processing hypothesis (1975)?

Answer 13

Shallow –e.g. how many letters does this word have?
Intermediate –e.g. does this word rhyme with ‘treat’?
Deep semantic – e.g. Truck. Do you find these in a city?

Question 14

What three forms of semantic organisation have been shown to enhance memory?

Answer 14

1. Categorisation at encoding
2. When more elaboration is encouraged during encoding
3. When material is self-relevant (e.g. instead of parsing adjectives into positive and negative, separate them into those that do/don’t apply to you)

Question 15

Does depth of processing (Craik and Lockhart, 1975) explain how memory works?

Answer 15

No, it just describes what kind of encoding encourages more long-lasting memories. It doesn’t tell us how and in what forms memories are stored.

Question 16

In what 2 ways do we retrieve information from the LTM?

Answer 16

1. Explicit voluntary retrieval eg memory search

2. Memory activated by retrieval cue: familiar; “just know” e.g. know name of friend upon seeing him

Question 17

What are the two standard ways of measuring memory –and which usually results in better performance?

Answer 17

1. Free recall task: report items from earlier study episode.
2. Recognition task: Select previously studied items from mixture of old and new items.

Recognition is easier.

Question 18

Why do recognition tasks result in better results than recall?

Answer 18

Recall task do not provide any (explicit) cues (only possibility that one item will trigger memory of another). Recognition tasks provides a cue (the studied item) which can activate (“prime”) memory network.

Question 19

What was Godden and Baddeley’s 1975 diving experiment – and what did it demonstrate about memory?

Answer 19

Participants learned words either on land or
20 feet underwater; and were asked to later recall the words on land or underwater.

Recall was best when the contexts matched.
Recognition was unaffected by retrieval context, only by learning context (easier to learn on land).

It demonstrated that the memory network can be primed by the study context. But the influence of context on retrieval depends on type of retrieval task.

Question 20

What is the encoding specificity principle (Tulving & Thomson, 1973)?

Answer 20

What we encode in memory is specific to the context in which we encode it. The specificity of that encoding means retrieval will be influenced by extent to which it matches that specific retrieval context.

Question 21

What is the ‘Transfer Appropriate Processing’ principle (Morris, Bransford & Franks, 1977)?

Answer 21

Retrieval is best when there is a match between the

processes required at encoding and retrieval.

Question 22

In what way does the Transfer-Appropriate Processing principle undercut the levels of processing model?

Answer 22

Memory performance depends not only on level of encoding, but ALSO on matching the learning and retrieval processes.

‘Shallow’ processing can yield BETTER memory when it matches the task context.

e.g. Study word: eagle
Deep - The ____ is the US national bird.
Shallow - rhymes with legal

THEN:

Either a standard recognition test, e.g. ‘Did you see the word before?’ Or a rhyming recognition test, e.g., ‘Was a word presented that rhymed with “regal”’?

Pps score higher on the rhyming recognition task –where processing is shallow but the task is transfer-appropriate.

Question 23

If people learn the sentence ‘The fish attacked the swimmer’, the word ‘shark’ is a better retrieval cue than ‘fish’. What does this demonstrate?

Answer 23

Transfer-appropriate processing. When words/sentences have been encoded semantically, the best cues relate to meaning, not surface form. ‘Shark’ is a better match for the GIST of the sentence.

Question 24

What’s the difference between implicit and explicit memory tasks?

Answer 24

Explicit memory tasks
Subjects explicitly told to remember items from
previous list. Intentional retrieval. Memory performance shows levels of processing and transfer-appropriate processing effects.

Implicit memory tasks
Subjects NOT told to try to remember, just to perform a task
e.g. fragment completion, stem completion, perceptual identification. Compare performance for old and new items to infer memory.

Question 25

What are 4 examples of memory processing not being ‘bottom-up’, thus contradicting the multi-store model of memory?

Answer 25

1. Semantic influences of context (LTM) on STM tasks
   – If a sentence makes sense, it’s easier to remember
   – Release from proactive interference*
2. Transfer-appropriate processing: same information ‘remembered’ differently in different tasks.
3. Information ‘retrieved’ in implicit memory tasks despite no rehearsal.
4. Implicit memory tasks ‘dissociate’ from explicit tasks ??.

Question 26

Describe Brown-Peterson (1959) study of proactive interference. What did it demonstrate about the effect of semantic information on STM?

Answer 26

Subjects are presented with 3 items (e.g. cherry, orange, grape) and asked to count backwards by 3s, then recall in correct order.
Recall deteriorates across successive trials from the
same category – e.g. if all fruit. – this is proactive interference.
But recall improves when items from a different semantic
category – e.g. flowers–are presented.

This shows that semantic information in LTM influences STM performance. It both creates proactive interference and releases from it.

Question 27

Describe the three modules in Anderson’s (1983) Adaptive Control of Thought model of memory.

Answer 27

1. DECLARATIVE MEMORY
   - Semantic memory
   - Episodic memory
2. PROCEDURAL MEMORY
   - Memory for how to do things, not verbalisable or available to conscious awareness. This is learnt through gradual, incremental experience, and does not require attention.
3. WORKING MEMORY
   The system in which incoming information is processed and integrated with existing declarative and procedural memories.