Memory and the Multi-Store Model

• These are study notes about memory and eyewitness testimony (EWT).

MSM - Atkinson & Shiffrin (1968)

• This is the Multi-Store Model of memory proposed by Atkinson and Shiffrin in 1968.
• According to this model, encoding is modality specific.
• The flow of information starts with sensory input, which is held briefly in the sensory register.
• Attention is required to transfer information from the sensory register to short-term memory (STM).
• Information in STM can be maintained through rehearsal.
• Elaborative rehearsal, which involves linking new information to existing information in long-term memory (LTM), aids in the transfer to LTM.
• STM has a limited capacity of 5-9 items and a duration of about 18 seconds.
• STM encoding is primarily acoustic.
• LTM has unlimited capacity and can last a lifetime, with semantic encoding.
• The model is considered too simplistic and inflexible.
• It is a linear model, suggesting information goes through stores in a set order.
• It is passive, holding information while it's passed on.
• The model assumes unitary stores, which has been challenged.

Sperling's Iconic Store

• Sperling's research focused on the iconic store (visual sensory memory).
• He determined the capacity to be approximately 4.2 bits.

Memory Definitions

• Coding refers to the format in which information is stored in different memory stores.
• Capacity refers to the amount of information that can be held in a store.
• Duration refers to how long information can be held in memory.

Sperling Study

• Aim: To investigate the capacity of visual sensory memory.
• Method: Participants were briefly shown 12 letters for 50 milliseconds.
• A tone indicated which row of letters participants should recall.
• Results: Participants could recall 3-4 letters from any row, suggesting all 12 letters were available in visual sensory memory briefly.
• Pro: Tight control, allowing cause and effect conclusions, high internal validity.
• Con: Low ecological validity and mundane realism.

Miller Study

• Aim: To investigate the capacity of short-term memory (STM).
• Method: Review existing research.
• Results: Average recall was approximately 7 +/- 2 items.
• STM capacity could be increased by grouping items into larger chunks.
• Pro: Supported by research of Jacob (1887), who found that digit span average recall was 7.3-9.3.
• Con: Low mundane realism, as recalling digits is not an everyday task.

Baddeley Study

• Aim: To investigate how information is coded in STM and LTM.
• Method: Participants were given word lists and asked to recall them either immediately (STM) or after a delay (LTM).
• Conditions included acoustically similar words, acoustically dissimilar words, semantically similar words, and semantically dissimilar words.
• Results: STM recall was worse for acoustically similar words, while LTM recall was worse for semantically similar words.
• Conclusion: STM codes acoustically, and LTM codes semantically.
• Con: Recalling word lists is not an everyday task, limiting ecological validity.
• Lacks mundane realism, as the words had little meaning to participants, limiting the representativeness.

Peterson & Peterson Study

• Aim: To investigate the duration of STM.
• Method: Participants were shown trigrams (three-consonant syllables) and asked to recall them after intervals of 3, 6, 9, 12, 15, or 18 seconds.
• To prevent maintenance rehearsal, participants counted backwards in threes from a specified number during the interval.
• Results: Most participants could recall trigrams after 3 seconds, but recall decreased significantly after 18 seconds.
• Duration of STM: Approximately 18 seconds.
• Pro: Controlled for interference, high internal validity.
• Con: Questionable whether recalling trigrams reflects everyday memory.
• Low mundane realism as most trigrams have no use in daily life.

Bahrick et al. Study

• Aim: To investigate the duration of LTM.
• Method: Participants were tested on their ability to recognize photos and names of their high school classmates.
• Results: After 15 years, photo recognition was 90% accurate, and name recognition was 60% accurate.
• After 48 years, photo recognition was 80% accurate and name recognition 30%.
• Duration of LTM: Possibly unlimited.
• Pro: Realistic, as it tested memory for meaningful information.
• Recalling names from lab.

Evaluating MSM

• Maintenance rehearsal is an important part of the model.
• Glanzer and Cunitz (1966) presented word lists and found better recall for the beginning and end of the list, showing primary and recency effects.
• Easy to understand, the model provides a foundation for later memory research.
• The model proposes unitary stores versus multiple stores.
• Criticisms include findings of visual and acoustic elements.
• The model is too simplistic - the STM store is not unitary.
• This criticism is supported by Baddeley and Hitch, 1974 - where the STM store has separate components such as the central executive.
• Shallice and Warrington found that the STM has distinct visual and acoustic elements.
• It is difficult to measure capacity.
• Jacobs (1887) researched capacity and found participants could recall around 7 letters/digits.
• Miller suggests the capacity is improved via chunking.
• Multiple LTM types exist.
• Tulving proposes multiple LTM types – episodic, semantic, procedural – when MSM says it’s unitary.
• Clive Wearing had a normal STM but damaged LTM after brain damage, challenging the unitary store assumption.
• Frost said Long-Term Recall is related to a Visual/Semantic sense.
• Nelson & Rothbart found acoustic encoding in LTM, when the MSM model says semantic.
• Flashbulb memories are embedded in LTM but don’t rely on rehearsal.

Types of LTM - Endel Tulving.

• LTM has three types: episodic, semantic, and procedural.
• The system of LTM is not anatomically unified but communicating.

Episodic memory

• Explicit/declarative memory of events in our lives, time-stamped.
• Associated with the hippocampus and prefrontal cortex.

Semantic memory

• Explicit/declarative memory of world knowledge, not time-stamped.
• Associated with the temporal lobe and prefrontal cortex.

Procedural memory

• Implicit/non-declarative memory of motor skills.
• Difficult to explain verbally, resistant to forgetting.
• Associated with the cerebellum and motor cortex.
• There are real world applications of understanding the different types of long-term memory.
• If one type of LTM is more affected than others, intervention can improve that type of memory.
• For example, Belleville training helped older adults to become better on episodic memory tests.

Evaluation on the Tulving LTM Model

• Neuroimaging has demonstrated that different brain activity occurs when we use different types of LTM.
• CW & HM both suffered amnesia but retained semantic and procedural memory, supporting that different types of LTM exist.
• However, it is difficult to judge how good or bad the memory was, making judging difficult.
• The weakness is that its hard to establish the location for each, as Tuling linked episodic memory to the left prefrontal cortex while Buckner and Peterson linked to the right prefrontal cortex.

Working Memory Model: Baddeley & Hitch (1974)

• This model is an alternative to the Multi-Store Model, suggesting that STM is not unitary but has several components.
• The model doesn't discuss LTM.
• Components include the central executive, phonological loop, episodic buffer, and visuo-spatial sketchpad.
• The central executive directs attention and gives priority to tasks, allocated resources.
• The phonological loop is a slave system used for auditory information, with an articulatory control system (inner voice) and phonological store (inner ear).
• The visuo-spatial sketchpad is for spatial and visual coding.
• The episodic buffer has a limited capacity and integrates information from different sources.

Research Support of Working Memory

• The working memory model is supported by:
• Shallice & Warrington (1970) found that KF had poor verbal STM but normal visuospatial - this meant that visuospatial was separate.
• Supporting by Baddeley for word length. This meant that Ppt can hold number of teams that can be said in 2 seconds.
• The capacity of the Central Executive is supported by Hunt
• dual performance. Baddeley found that visual/verbal tasks perform the same.
• This also supports distinct Slave Systems (VSS & PL).

Evaluating WMM Model

• Strengths include that research supports with task studies, WM is more possible than MSM which uses memory as short-term, and there is case study support with PL and VSS
• Weaknesses include WM’s only concern is Short-Term Memory, it’s an incomplete model because we do not know much about CE and WM doesn’t state how the working memory works the same for all.

Explanations of Forgetting - 2 Theories

Interference theory of forgetting

• Assumes one set of learning interferes with another, impairing recall from LTM.
• Most likely to occur when memories are similar.
• Proactive interference occurs when an older memory interferes with the ability to retrieve a newer one.
• Retroactive interference occurs when a newer memory interferes with the ability to retrieve an older one.
• A support of this includes Jenkins & Dallenbach (1924).

Retrieval Failure Theory

• Occurs when a memory is available but not accessible, with a lack of necessary cues to access the memory.
• A cue is a trigger encoded during learning that helps access the memory, which can be external or internal.
• Supported by research such Godden & Baddeley (1975) and Carter & Cassaday (1998).

Encoding Specificity Principle

• Tulving (1983) suggests if cues available at encoding and retrieval are different, forgetting will occur.

Anxiety

• Anxiety is a state of emotional and physical arousal.
• Physical changes can include increased heart rate and sweatiness.

Negative Effect on EWT

• Anxiety can create psychological arousal in the body.
• Reduced recall of details.
• The presence of a weapon creates anxiety, leading to a focus on the weapon.
• This causes a reduction in other details - called weapon focus effect.

Testing the effect of a Weapon

• Anxiety and weapon testing is supported by Johnson & Scott.

Post-Event Discussion

• Involves discussion between multiple witnesses of an event, which may affect the accuracy of recall.
• Post-event discussion is seen in Gabbert et al (2003).

Leading Questions and EWT

• 75% of innocent convictions from EWT are from DNA testing.
• Factors affecting EWT include misleading information and leading questions.
• Leading questions involve the way and questions are being asked in.

Leading Questions: Loftus & Palmer (1974)

• Aim: To investigate the impact of leading questions on estimates of car speed.
• The investigation involved questions/language for estimates of car speed.

Improving the Accuracy of the Interviews

• Improving accuracy involves cognitive interviews to reduce the influence of EWT (Eyewitness Testimony).