Memory 3: Learning Flashcards
(23 cards)
How does the WMM fail to explain semantic chunking?
- Words in a sentence are semantically chunked (dividing many words into chunks based on similar meaning) via LTM semantics, and many more items can be remembered in this case (16)
- e.g. Patient PV has a digit span of 2, but a sentence span of 6, and can verify long meaningful sentences
Research has shown that semantic chunking can happen in absence of LTM
Where in WM does semantic chunking take place?
Can semantic chunking happen independent of LTM?
Yes
Amnesic patients show supra-span prose recall in absence of LTM (Baddeley & Wilson, 2002)
They had very good immediate recall of prose using semantic chunking in sentences, but very poor delayed recall - shows no role of the LTM
How does the WMM fail to explain visual similarity effects in verbal recall?
- Small advantage for visually dissimilar words (guy, sigh, lie) over visually similar words (fly, dry, cry) in verbal serial recall
- Indicates both verbal and visual coding for visually presented words
- However 1974 model shows verbal and visual as separate stores
So how is verbal and visual combined
What revision was later made to the multicomponent (WM) model?
Episodic buffer - storage system - holds chunks of multidimensional (visual, verbal, semantic) information via connections to the sketchpad, loop, central executive and LTM, and binds multidimensional information together to an integrated representation that is accessible to conscious awareness.
How does the presence of the episodic buffer answer these three questions?
1) How does WM interact with LTM?
2) Where does semantic chunking take place?
3) How is verbal and visual info combined?
1) How does WM interact with LTM?
- Loops of rehearsal and retrieval between buffer and LTM - generate and activate long term info
2) Where does semantic chunking take place?
- Buffer capacity - 4 chunks of multi-dimensional information
3) How is verbal and visual info combined?
- Buffer has direct links with the subsystems and thus holds integrated multimodal representations
What happens when a reasoning task and a digit repetition task are done concurrently?
Digit load slows down reasoning task, but does not increase errors
- Reasoning time increases as a function of digit load, but error rates in the reasoning task remain constant across digit load - Digital load (concurrent STM task) slows down reasoning speed - Digit load slows down reasoning speed, but reasoning is absolutely possible despite digit load - Information retention (digits) and information manipulation (reasoning) work in parallel
Why is WMM better than a multistore model?
It replaces a unitary, purely storage-focused, memory component with a dynamic mental workspace model that explains how different storage and processing components interact, to allow for online information manipulation, needed to perform complex cognitive tasks (e.g., learning, reasoning, mental imagery).
What is Ebbinghaus’s total time hypothesis?
The more time you invest for learning, the greater is the amount of material you retain
How did Ebbinghaus conduct his experiment on the rate of learning?
He learned lists of 16 nonsense sets of 3 syllables, repeated recitation 8, 16, 24, 32, 42, 53 or 64 times on Day 1, and on Day 2 - recalled after 24 hours
What is Ericsson et al (1993) critique of the total time hypothesis? (violinist experiment)
Time invested is not the only contributor to memory
- 10 best student violinists, 10 good student violinists - Ppts keep activity diaries (duration of musical activities during one week - The percentage of time spent on practise (time alone with the violin) does not differ between the best and the good students (~24.3 hours invested per week in both groups) - Time invested IS NOT everything
What did Ericsson argue matters more to memory than time invested?
Deliberate practise
Extensive practise enhances performance, but this eventually levels out.
To become a pro and to further improve performance, it takes deliberate practise (e.g., focus on weaknesses, development of new skills) as opposed to merely engaging in many hours of activities.
Maths on Ebbinghaus:
2.5 syllables per second, can calculate amount of time each different amount of repetition took on Days 1 and 2
Doing more practice on Day 1 results in 14 minutes total practice (7 mins each day), doing less practice on Day 1 results in 21 minutes total practice (1 min on Day 1, 20 minutes on Day 2)
What does this show is more important than total time for learning?
Disproportionate saving
- Missing learning on Day 1 means proportionally much more learning is required on Day 2 - time spent on Day 1 gives a disproportionate saving on re-learning on Day 2
This is the distributed practice effect
What is the distributed practice effect?
Evenly distributed learning results in more (time) efficient learning than cramming most of the practise into the day of recall
Landauer & Bjork (1978) - how did they think learning and testing sequences should change for optimum memory?
- Learning and testing sequences (for the same material) should alternate.
Items should initially be tested after a short delay after learning
Then pre-test delays should gradually increase across subsequent trials.
What did Pashler et al. (2007) think the interstudy interval should be?
- Optimum: Interstudy interval should be 10-20% of the retention interval, e.g., 1-2 days learning interval for testing after 10 days.
Bekerian and Baddeley (1980) - is retrieval of memory enahnced purely by repetition?
When asking participants to recall a very repetitive BBC advert, what did they find?
- When the BBC changed their wavelengths for tuning into radio they advertised this heavily on the radio. Very repetitive advert.
- How successful was this advertisement?
- 50 participants were tested, almost all were aware of the change, only about 25% even tried to recall the numerical frequencies of the new wavelengths, and they were at chance level.
- Simple repetition and mere exposure, with no attempt by the learner to organise or meaningfully integrate the material, does not always lead to learning.
- e.g. try and remember the sign on your favourite pub
- For explicit learning - you have to engage
What are the two combined principles that explain expanding retrieval? (Second one has a and b)
- Distribution of practise/ spacing effect
- (a) The testing effect
(b) The generation effect
Testing effect
Roediger & Karpicke (2006)
They did an experiment with three conditions - what were they?
- Learnt a prose passage
- 3 conditions: repeated study (4x reading), single test (3x reading, 1x testing), and repeated test (1x reading, 3x testing)
- Free recall after 5 mins and 1 week
Roediger & Karpicke (2006)
Which learning condition produced the best recall after 5 mins vs after 1 week?
- After 5 minutes, recall performance is best for the study only condition, followed by the single test and repeated test condition.
- After 1 week, recall performance is best for repeated test, followed by single test and study only.
- Long term - testing is better - the more you test yourself on what you have learned, the longer you will remember it
What does long term repeated testing lead to?
- Repeated testing leads to enhanced organisational processes,
i.e., enhanced organisation of the material and enhanced organisation of recall per se (Zaromb & Roediger, 2010).- Repeated testing supports the use of more effective mediators during learning,
i.e., meaningful links are established during learning which are updated due to experienced retrieval failure (Pyc & Rawson, 2010).
- Repeated testing supports the use of more effective mediators during learning,
What is the generation effect? (Metcalf and Kornell, 2007)
Learnt English-French vocabulary
- Conditions - simultaneous presentation of both words (passive learning), and presentation of the French word after a delay (generation)
- Cued recall - English presented, French recalled
- More words were correctly recalled in the generation as compared to the passive learning condition.
- Meaning, material that you actively recall from memory (self-generated), is remembered better than material you are passively presented with
- When you self-test, you generate maximally
- Feedback is needed - repeated study/testing alone means that errors made at the beginning tend to persist - this is eliminated when feedback is given
What are the six empirically confirmed study tips?
- Practise deliberately, i.e., focus on weaknesses, development of new skills.
- Distribute your studying over time.
- Engage, i.e., organise and integrate new material, if you want to learn explicitly.
- Test yourself repeatedly on what you have studied.
- Actively retrieve answers from memory, don’t just consume.
- Give yourself feedback, i.e., keep correcting yourself.
Gurung (2005) - psychology students did a questionnaire after end-of-year exams
Investigated whether there were correlations between specific study techniques, effects of distractors, class attendance, total hours studied, and final exam performance.
What did they find?
- More hours students spent studying, memorising their notes, making up examples (e.g., to incorporate material into everyday life), reading the text book, reading their notes, using mnemonics (e.g., visual imagery), and self-testing their knowledge, the higher their exam scores
- Students who missed class a lot, and who had music/TV on, responded to emails/texts, had friends around while studying had lower exam scores
- Most frequently used - reading notes, reading notes, using mnemonics (all positively correlated with exam scores)
- Frequently used - rewriting notes, reviewing highlighted material, reviewing figures and tables in the text (no correlations)
- Least frequently used technique and STRONGEST predictor for a good mark - self testing
