Two Key Studies of Reconstructive Memory: Bartlett (1932) & Loftus & Palmer (1974) (SL IB Psychology)

Aim: To investigate the effect of cultural schemas on reconstructive memory

Participants: 20 male undergraduate students from the University of Cambridge in the UK

Procedure: Bartlett instigated a procedure known as serial reproduction, in which one participant read the story then reproduced it in writing; this was then read to a second person who then wrote his own memory of the story which was then read to a third person who then produced his own version of the story and so on.

Results: Bartlett found that the resulting stories bore little similarity to the original Native American folk tale. The changes made by the participants included:

• Omission: Key details of the story were ignored or missed out, particularly unfamiliar or unpleasant details such as a contorted face or black coming out of a mouth. Participants even omitted the key idea that ghosts were fighting which is surprising as this is the title of the story. Ghosts were soon dropped from the re-telling of the story as they do not fit with the way that adult males see the world, particularly in relation to war; details such as a contorted face were omitted as they may have caused unpleasant memories.
• Assimilation and sharpening: Story details were changed to suit the participants’ own cultural schemas e.g. ‘canoes’ became ‘boats’; ‘paddling’ became ‘rowing’. Details such as the spirit wound were re-interpreted as a flesh wound with words such as ‘therefore’ and ‘because’ inserted to explain the events.
• Levelling: The story became shorter - the original story was approximately 350 words and the participants’ version was around 180 words

Conclusion: Cultural schemas contribute to the reconstructive nature of memory i.e. memory is not a passive state in which events are recorded like a camera would record them, instead memory is an active process in which pre-existing information and expectations may interfere with the accuracy and reliability of the memory

Evaluation of Bartlett (1932)

Strengths

• Bartlett’s study was one of the first pieces of research to highlight the role of schema in reconstructive memory e.g. two people who witness the same event may give very different accounts of what they have seen
• Understanding the ways in which schemas may interfere with accurate recall of events has good application to educational settings in terms of how learning takes place and to the criminal justice system in terms of eye-witness testimony (see Loftus & Palmer 1972 below and the Revision Note on Eye Witness Testimony)

Limitations

• This is very dated research: university students in the UK are much more aware of wider multi-cultural issues today than they were in the 1930s which means that the results may lack temporal validity
• It is unclear as to whether the memory distortions were the product of schematic interference or to other factors such as poor overall memory, lack of attention, personal learning styles (some people are visual learners for example so an aural task would not suit them as much as viewing a cartoon of the story)

Key terms:

• Assimilation
• Levelling
• Reconstructive memory

Aim: To investigate the effect of leading questions on eyewitness testimony (EWT)

Participants: 45 undergraduate students from the University of Washington, USA for Experiment 1; 150 participants from the same university for Experiment 2

Procedure: Two lab experiments which used an independent measures design for both Experiment 1 and Experiment 2

• Experiment 1: Participants were shown seven film clips of traffic accidents. After each film they filled in a questionnaire based on what they had witnessed about the accident – the questionnaire included several ‘filler’ questions and a critical question
• The critical question (independent variable) was: ‘About how fast were the cars going when they smashed/hit/bumped/collided/contacted each other?’ Each participant was in one of the five conditions i.e. each participant was asked only one of the critical questions containing only one of the five verbs. Participants had to estimate the speed in miles per hour 
• Experiment 2: 150 participants divided into three groups of 50 each. All participants watched a one-minute film of a multiple-car accident. They then answered some questions about the film
• The critical question was, ‘How fast were the cars going when they hit/smashed each other?’ Each participant was randomly allocated to either the ‘smashed’, ‘hit’ or control condition. The control group were not asked any questions about the speed of the cars

The participants were asked to return a week later. They were asked several questions about the accident in the film. The critical question was, ‘Did you see any broken glass?’ with the response being ‘yes’ or ‘no’.  There was not, in fact, any broken glass in the film

Results: 
Experiment 1: Participants in the ‘smashed’ condition estimated the highest speed out of all the five conditions at 40.8 mph; participants in the ‘contacted’ condition estimated the lowest speed out of all the five conditions at 31.8 mph

Experiment 2: 43 participants in the ‘Smashed’ condition reported having seen broken glass as opposed to 7 participants reporting seeing broken glass in the ‘Hit’ condition

Conclusion: Leading questions may lead to unreliable EWT by providing information after the event

Evaluation of Loftus & Palmer (1974)

Strengths

• This research has huge implications for the ways in which EWTs should be questioned hence it has great application to the wider world
• The standardised procedure and control of variables make this study easy to replicate which increases its reliability

Weaknesses

• Watching recorded footage of a traffic accident is not the same as experiencing the event in real life so the study lacks ecological validity
• The participants might have been prone to response bias - i.e. the emotive quality of the words may have prompted the participants to think that a higher or lower speed estimate was expected of them (e.g. ‘smashed’ sounds like it should be given a high estimate)

Key terms:

• Eyewitness testimony
• Information after the event
• Response bias