All key memory studies

• Baddeley + Hitch 1976 (MSM + WMM)

  • Aim: investigate STM's functions
  • Sample: 92 students
  • Method: lab study + IM design
  • Procedure: dual-task paradigm
  • Describe letter pairs while doing sound-related activities, e.g. repeat 'the'/'1, 2, 3, 4, 5, 6'/, digit strings
• Baddeley + Hitch 1976 (MSM + WMM)

  • Findings: reasoning task only impaired during string of out-loud digits
  • Conclusions: STM = different components + contradicts MSM
  • Strengths: lab experiment = high internal validity -> replicable = reliable + IM design = avoids order effects
  • Weaknesses: lab = too artificial = low ecological validity + IM = possible ppt variables
• Peterson + Peterson 1959 (MSM)

  • Aim: investigate STM's duration
  • Sample: 24 psychology students
  • Method: lab experiment using trigrams
  • Procedure: ppts counted backwards in 3s while shown trigrams; asked to repeat trigrams after 3/6/9/12/16/18s
• Peterson + Peterson 1959 (MSM)

  • Findings: recall declined with time
  • 3s = 80%
  • 6s = 50%
  • 18s = 10%
  • Conclusions: info lost from STM within 20s without rehearsal
  • Strengths: lab experiment = good control -> replicable = reliable
  • Weaknesses: small sample, sample bias (all psychology students) = results can't be generalised + artificial task = low ecological validity
• Baddeley 1966 (MSM)

  • Aim: investigate STM + LTM's coding
  • Sample: 72 psychology students
  • Method: lab study, IM design for IV 1 + RM design for IV 2
  • Procedure:
  • IV 1 = ppts learned acoustically similar/different word list
  • IV 2 = ppts learned semantically similar/different word list
  • Recalled immediately for STM + after time delay for LTM
• Baddeley 1966 (MSM)

  • Findings:
  • IV 1: most errors for STM = acoustically similar
  • IV 2: most errors for LTM = semantically similar
  • Type of error depends on type of list + time delay
  • Conclusions: STM codes acoustically + LTM codes semantically
  • Strengths: lab experiment = high control, IV/DV operationalised well + reliable (extraneous variables controlled)
  • Weaknesses: artificial task + non-naturalistic materials = low ecological validity + sample bias (100% psychology students)
• Bahrick et al. 1975 (MSM)

  • Aim: investigate LTM's duration
  • Sample: 392 US ppts who graduated 1-48 years prior
  • Method: lab experiment using naturalistic materials (old photos + classmates' names)
  • Procedure: ppts tested on recognition of classmates' faces/names, matching names to pictures + free recall of names
• Bahrick et al. 1975 (MSM)

  • Findings: name recognition recall very high
  • 34 years = 90%
  • 48 years = 80%
  • Free recall: 30 years = 30%
  • Conclusions: LTM = durable + pictures easier to recall than names
  • Strengths: real-life materials relating to everyday life = high ecological validity + other research support = reliable
  • Weaknesses: sample bias (100% US) = reliability lowered, harder to control (rehearsal, e.g. seeing classmates IRL = task easier)
• Milner's Patient HM 1957 (MSM + types of LTM)

  • Aim: investigate whole of LTM
  • Sample: man who had a lobotomy to treat trauma-induced seizure disorder
  • Method: case study
  • Procedure: studies testing STM/LTM + studies on HM's brain
• Milner's Patient HM 1957 (MSM + types of LTM)

  • Findings: contradicts MSM
  • HM = no new memories post-operation but could form new motor skills
  • Working non-declarative memory, defective declarative memory
  • Conclusions: new memories = medial temporal lobe + LTM has different stores -> MSM too simplistic
  • Strengths: idiographic research = in-depth + variety of methods = reliable
  • Weaknesses: too subjective = low replicability + surgeries/studies too experimental = unethical
• Shallice + Warrington 1970: KF (WMM)

  • Aim: investigate relationship between STM + LTM when STM is impaired
  • Sample: KF (man with brain damage after accident)
  • Method: case study
  • Procedure: range of qualitative measures (interviews, observations, etc) for longitudinal study
• Shallice + Warrington 1970: KF (WMM)

  • Findings: KF found it difficult to process + recall auditory stimuli but easy to recall visual stimuli
  • Conclusions: STM has multiple stores that code different types of info
  • Strengths: rigorous study = reliable + case study = high ecological validity
  • Weaknesses: small, unique sample = not generalisable + researchers' close focus = potential investigator bias
• Christianson + Hubinette 1993 (anxiety's effect on EWT accuracy)

  • Aim: investigate anxiety's effect on EWT accuracy
  • Sample: witnesses to 22 bank robberies
  • Procedure: witnesses interviewed about robberies
  • Direct victims = higher recall of robbers' appearances than onlookers
  • Conclusion: increased anxiety = increased EWT accuracy
• Yuville + Cutshall 1986 (anxiety's effect on EWT accuracy)

  • Aim: investigate anxiety's effect on EWT accuracy
  • Sample: witnesses to crime involving thief with gun + shoot-out
  • Procedure: witnesses interviewed about crime
  • Findings: central witnesses = higher recall than peripheral witnesses
  • Conclusions: increased anxiety = increased EWT accuracy
• Johnson + Scott 1976 (anxiety's effect on EWT accuracy)

  • Aim: investigate anxiety's effect on EWT accuracy
  • Method: lab experiment + IM design
  • Procedure:
  • IV 1: man walks out of room after conversation holding a pen
  • IV 2: man walks out of room after argument holding a bloody paper knife
  • Ppts had to identify him afterwards
  • Findings: IV 2 = worse recall than IV 1
  • Conclusions: increased anxiety = decreased EWT accuracy
• Yerkes-Dodson Law (anxiety's effect on EWT)


  High/low anxiety = low accuracy of EWT

  Moderate anxiety = high accuracy of EWT
• Loftus + Palmer 1974 - STUDY 1 (factors affecting EWT)

  • Aim: investigate how post-event leading questions affect EWT accuracy
  • Sample: 45 students
  • Method: lab experiment + IM design
  • Procedure:
  • Ppts watch series of video clips -> one is of a car crash
  • Ppts given questionnaire about clips -> one Q is "how fast were the cars going when they contacted/bumped/hit/smashed/collided (into) each other?"
• Loftus + Palmer 1974 - STUDY 1 (factors affecting EWT accuracy)

  • Findings: 'smashed' = highest speed estimates + 'contacted' = lowest speed estimates
  • Conclusions: leading questions either impact accuracy of recall or just cause a response-bias
• Loftus + Palmer 1974 - STUDY 2 (factors affecting EWT accuracy)

  • Aim: investigate how post-event leading questions affect EWT accuracy
  • Sample: 150 students
  • Method: lab experiment + IM design
  • Procedure:
  • Ppts shown short video of multi-vehicle crash
  • Ppts asked "how fast were the cars going when they hit/smashed into each other?"
  • One week later: ppts asked if they saw broken glass (there was none in the video)
• Loftus + Palmer 1974 - STUDY 2 (factors affecting EWT accuracy)

  • Findings: 'smashed' = more reports of broken glass than 'hit'
  • Conclusions: leading questions cause altered memories, not just response-bias
• Gabbert et al. 2003 (factors affecting EWT accuracy)

  • Aim: investigate effect of post-event discussion on EWT accuracy
  • Sample: 60 students from Uni of Aberdeen + 60 older people from local community
  • Method: lab experiment + IM design
  • Procedure:
  • Ppts shown video of girl stealing money from a wallet
  • Experimental group discussed whether she stole the money with 'co-witness' -> both told they watched the same video but one person actually watched a different one with more detail
  • They both filled in questionnaire about the video afterwards
  • Control group filled in questionnaire without discussion
• Gabbert et al. 2003 (factors affecting EWT accuracy)

  • Findings: 71% of experimental witnesses recalled details they hadn't seen + 60% said she was guilty despite not actually seeing her commit the crime
  • Conclusions: people = prone to 'memory conformity' -> post-event discussion = altered memory/reduced EWT accuracy