Eyewitness Memory

Document Sample
Eyewitness Memory Powered By Docstoc
					                                                                         Spiro (1980)
                                                          • Suggested that when memory changes to
         Eyewitness Memory                                  accommodate new information added after
                                                            the original experience:
                                                            (1) the chance of such changes increases with
          Chapter 15, pages 334-338                           time
                                                            (2) the confidence the subject has that the
                                                              memory is accurate will be as high or higher
                   Erin Pieper
                                                              than for accurate memories
                                                            (3) there is no way to distinguish inaccurate
                                                              memories from accurate memories




  Supporting Research (Loftus)                                 (Loftus and Palmer, 1974)
• Basic Paradigm:                                         • Subjects see a video depicting a car accident
  – subjects see a series of slides or a film depicting   • Asked, “How fast were the cars going when
    a car accident, robbery, or some other                  they________ each other?”
    naturalistic eyewitness event
                                                          • Varied the intensity of the verb that
  – asked a series of questions
                                                            described the collision: smashed, collided,
  – manipulate the way one question is asked to
    determine effects on recall
                                                            bumped, hit, and contacted




   Results of Loftus and Palmer                              Loftus and Palmer Continued
                                                          • Follow-up question asked whether there
                                                            was broken glass
                                                            – “Smashed” condition- 1/3 said yes
                                                            – “Contacted” condition- 1/10 said yes
                                                          • There actually was no broken glass
                                                          • Conclude that the verb used led subjects to
                                                            reconstruct their memory to be consistent
                                                            with the inplied speed of the verb




                                                                                                             1
  Loftus, Miller, & Burns, 1978                   Loftus, Miller, & Burns, 1978
• Words used in asking questions can change     • Asked either a consistent or misleading
  the answers                                     question: “Did another car pass the red
• Subjects watched a sequence depicting a car     Datsun while it was stopped at the _____
  (a red Datsun) that makes a turn a hits a       sign?”
  pedestrian                                    • Consistent- i.e. saw a stop sign and the
  – 50% saw a STOP sign at the intersection       question mentioned a stop sign
  – 50% saw a YIELD sign at the intersection    • Misleading- i.e. saw a stop sign and the
                                                  question mentioned a yield sign




  Loftus, Miller, & Burns, 1978                     Loftus et al., 1978: Results
• 20 minutes later subjects received an                            Saw Stop        Saw Yield
  recognition test in which they were shown                        Sign            Sign
  two photographs:                                Question         (Consistent)    (Misleading)
• One had a stop sign and one had a yield         Mentioned        75% Recall      40% Recall
  sign                                            Stop Sign        Accuracy        Accuracy
• Asked which one they had seen                   Question         (Misleading) (Consistent)
                                                  Mentioned        40% Recall 75% Recall
                                                  Yield Sign       Accuracy     Accuracy




    Loftus et al., 1978: Results                    Loftus et al., 1978: Results
                  Saw Stop       Saw Yield      • When subjects were tested 2 weeks after the
                  Sign           Sign             original incident the effect was greater:
  Question        (Consistent)   (Misleading)     – only 20% accuracy in the misleading condition
  Mentioned       75% Recall     40% Recall         (Subjects far more likely to recall the planted
  Stop Sign       Accuracy       Accuracy           information than the original information)
                                                • Even when told of the experimental design
  Question        (Misleading) (Consistent)
  Mentioned       40% Recall 75% Recall           90% of the misled subjects believe that they
  Yield Sign      Accuracy     Accuracy           received consistent information which they
                                                  reported accurately




                                                                                                      2
                                                                                                    Biased Guessing Account
                                       Controversy
                                                                                                  (McCloskey & Zaragoza, 1985)
    • Is the information permanently lost, or is it                                              • Emphasizes the fact that the misleading
      retained but inaccessible?                                                                   information is more recent than the original
    • Impairment Account (Loftus):                                                                 information
           – memory overwritten by new information                                               • Experiments involve a forced choice
    • Biased-Guessing Account (McCloskey &                                                         between the original and misleading
      Zaragoza)                                                                                    information
           – memory for the original information is still
             retained




         Biased Guessing Account
       (McCloskey & Zaragoza, 1985)                                                               Summary of the Two Accounts
                        C o n trol G rou p

                                                                                                 • Impairment Account (Loftus):
R e m e m b er an d P ic k              F org et a n d G u e ss
                                                                                                   – the control subjects should be more accurate
                                                                                                     then the experimental subjects because their
         O rig in a l                O rig in a l         M is le ad in g
           50%                         25%                    25%                = 75% Correct       memory for the original memory was not
                                                                                                     overwritten
                          M isle d G ro u p
                                                                                                 • Biased-Guessing Account (McCloskey &
 R e m e m b e r a n d P ick                  F o rg e t a n d G u e ss                            Zaragoza)
                                                                                                   – performance should be nearly equivalent for the
O rig in a l      M is le a d in g         O rig in a l       M is le a d in g                       two groups (memory not overwritten)
  50%                 30%                    10%                  10%            = 60% Correct




       McCloskey and Zaragoza, 1985                                                                Emotional Intensity of Events
    • Designed an experiment to distinguish between                                              • Criticism: Research events are not as
      these two accounts                                                                           emotionally intense or arousing as real
    • Rather than a forced choice between the original                                             events
      and the misleading item, used the original and a
      novel item
                                                                                                 • Wagenaar & Groeneweg (1990) studied
                                                                                                   memory of concentration camp survivors
    • Performance by the control and experimental
      groups did not differ                                                                      • Conclusion: the intensity of emotion at the
    • Does not dispute the basic effect that people are                                            encoding of the information is no guarantee
      very likely to incorporate and report extraneous                                             for accurate eyewitness testimony
      information




                                                                                                                                                       3
   Attribution Errors (Baddeley,
                                                                    Summary
               1990)
• Case in which a woman was raped and             (1) new information can interfere with memory of an
  identified the man who was on television at       original event
  the time of the assault as the attacker         (2) the chance of new information interfering with
                                                    information for the original event increases with
• When a memory is recollected, the source          time
  of the memory can be attributed either
                                                  (3) there is no way to distinguish inaccurate
  accurately or inaccurately                        memories from accurate memories without
• Even if all of the details are correct, an        objective evidence
  attribution error can result in fundamentally
  inaccurate information




                                                                                                        4