Reconstructive memory and the encoding specificity principle

Question 1

Evidence that memory is reconstructive:

Eyewitness testimony study. Showed that memory of an event can be interfered with by post-event questioning

45 students, shown short clips, and asked to write an account of the accident they had just seen. Also asked to answer some specific questions. Five conditions in the experiment, the only thing varied was the adjective to do with the speed of the vehicles: smashed/collided/bumped/hit/contacted. Asked to esimate how fast the cars were going when they **** eachother.

Results showed clear effect of adjective- e.g. smashed av esitmate was 41mph, contacted was 32mph.

2 possible explanations:

• distortion in memory of participant
• response-biase factors: participant is not sure of eteh exact speed, so adjusts etimate to fit in wiht expectations of the questioner.

Second experiment: similar proceedure. Three conditions: 50 asked how fast were the cars going when they hit, 50 asked how fast when they smashed, 50 not interrogated about speed. 1 week later, participants returned and without viewing the film again were askes series of questions about the accident- critical question was ‘did you see any broken glass?’. (there was no broken glass). ‘Smashed’ condition were significantly more likely to say yes, compared to hit and control.

L and P explained this with the reconstructive hypothesis- that 2 kinds of information go into a persons memory of an event: the info obtained from percieveing the event, and teh other info supplied after the event. Over time the info from these two sources may be integrated in a way that cannot tell what came from which source.

NB limitation: artificial situation, not how people normally witness events. ALso used all student participants- not representative of the general pop, especially not many tend to be experienced drivers.

Answer 1

Loftus and Palmer 1974

Question 2

Proposed the generate-recognise model

• Retrieval involves generating possible candidates for memory being sought, then testing whether these candidates are recognised as heaving been encountered recently.
• Can account for reliable performance differences seen between episodic memory tasks:
  
  • Free recall is typically worse than cued recall, which is generally worse than recognition
  • Free recall: generation phase has less info to guide it than when a cue is present
  • Recognition: fallible generation phase is bypassed altogether in favour of simple detection of whether stimulus word was encountered earlier.

Prediction: any item that can be recalled should also be correctly recognised, because recall involves a recognition stage:

Answer 2

Anderson and Bower 1972

Question 3

Tested and falsified key prediction of the generate-recognise model: that any item that can be recalled should also be correctly recognised, because recall involves a recognition stage:

4 phases:

1. Subjects presented with a list of weakly-associated word pairs (e.g. train-BLUE).
2. Then shown a word (e.g. sky) that was strongly associated to one of the target words, and asked to generate associated (CLOUD, BLUE, etc)
3. Recognition test for words participant had generated: was cloud or blue in the original list.
4. Cued recall test given (e.g. what word was paired with train in original list)

Contrary to generate-recognise model, words recalled in phase 4 were often not recognised in phase 3.

Then proposed the encoding specificity principle: = that the probability of retrieving an item from memory is related to the degree of overlap between the processing operations occurring at encoding and retrieval

Answer 3

Tulving & Thompson 1973

Question 4

Evidence for the encoding specificity principle.

Presented word-pairs that were either weakly or strongly associated. Tested recall with a cue that was weak or strong.

• Retrieval was much better when recall cue overlapped with encoding cue

Answer 4

Tulving & Thompson 1970

Question 5

Evidence for the importance of context in which info is encoded and retrieved, and the encoding specificity principle.

Asked scuba divers to learn words either on land or underwater, then tested recall either on land or under water.

• Retention was 50% higher when learning and recall took place in same context

Answer 5

Godden and Baddeley 1975

Question 6

Evidence that overlap in the mood participants were in during learning and recall strongly affected likihood of retrieval success.

Manipulated subjects moods, and whether or not observable retrieval cues were present.

• Subjects who learned and recalled in different moods had significantly greater decrements in recall than did subjects in the same moods
• The presence of observable retreival cues at recall overrode state-dependent recall. By manipulating the presence/absence of observable cues at recall, both the occurance and the erasure of the mood state dependency was demonstrated
• In a further investigation, mood state during learning and cued recall was also shown to affect performance in a third session under conditions of free recall.

Answer 6

Kenealy 1997

Question 7

Aimed to investigate how memory of a story is affected by previous knowledge.

• Used serial reproduction technique: particupants hear a story or see a drawing and are asked to reproduce it after a short time and then do so again repeatedly over a period of days, weeks, months or years.
• Told participants a Native American legend called The War of the Ghosts.
• Study participants were British; for them the story was filled with unknown names and concepts, and the manner in which the story was developed was also foreign.Therefore story was ideal to study how memory was reconstructed based on schema processing.

Results: Distorition: participants changed the story as they tried to remember it. Three patterns of distortion took place:

• Assimilation: story became more consistent with the participants’ own cultural expectations - that is, details were unconsciously changed to fit the norms of British culture.
• Levelling: story also became shorter with each retelling as participants omitted information which was seen as not important.
• Sharpening: participants also tended to change the order of the story in order to make sense of it using terms more familiar to the culture of the participants. They also added detail and/or emotions.

Participants overall remembered the main themes in the story but changed the unfamiliar elements to match their own cultural expectations so that the story remained a coherent whole although changed.

Lead researcher to the hypothesis that memory is reconstructive. Memories are not copies of experiences but rather reconstructions.

Limitations:

• Performed in a lab- lack of ecological validity
• Participants did not reciev estandardized instructions, so some of the distortions could be due to participant’s guessing or other demand characteristics.
  *

Answer 7

Bartlett 1932