Memory - Key studies

Peterson and Peterson 1959 - STM duration

Gave participants a trigram to remember. They were then given a 3 digit number and had to count backwards from this number in 3s aloud for varying periods of time. This prevented mental rehearsal of the trigram. After 3 seconds (of counting aloud), average recall was 80%. After 18 seconds, recall was 3%. Peterson x2 concluded duration of short term memory is around 18 seconds.

Baddeley 1966 - coding

Baddeley gave Ps 2 lists of words for them to learn. He found that when asked to recall them immediately (recalling from STM), they tended to do worse with acoustically similar words. But when asked to recall the lists 20 minutes later (LTM), they did worse with semantically similar words. Concluding STM is acoustically coded and LTM is semantically coded. Support for the multi-store model.

Bahrick et al. 1975 - duration of LTM

Used school yearbooks to assess the duration of long term memory. Recall was tested through photo recognition and free recall of their old classmates. Ps who left school less than 15 years ago averaged 90% accuracy in photo recognition. Ps who left school over 48 years ago, averaged 70% accuracy. Free recall was less accurate (60% for 15 years, dropped to 30% for 48 years). Also acts as supporting evidence for retrieval failure/cues.

Patient HM - MSM support

Had his hippocampus removed. He could not form any new long-term memories but could remember long term memories from before the surgery e.g. childhood friends. His short term memory was still intact though. Providing support for the multi-store model that LTM and STM are separate stores.

Baddeley and Hitch 1975 - dual tasks WMM

Lab experiments where Ps completed 2 tasks simultaneously. If both tasks used the same store of the working memory model, performance was worse. E.g. Tasks both using the phonological loop include repeating a word aloud whilst silently reading a story. He concluded that competition for attentional resources within the same component of the WMM inhibits recall.

Patient KF - WMM support

Suffered brain injury in a motorbike accident. KF had poor STM ability of auditory information but could process visual information normally. KF’s phonological loop had been damaged but his visuo-spatial sketchpad was intact.

McGeoch and McDonanld 1931 - interference

Ps learned a word list to 100% accuracy, they then had to learn new list. When Ps were asked to recall the original list, the more similar the 2nd list was to the original (e.g. synonyms vs numbers), the worse their recall was. This supports retroactive interference.

Godden and Baddeley 1975 - ESP

Studied deep-sea divers who were asked to learn a list of words either on land or underwater, they were then asked to recall either on land or underwater. They found recall of words was more accurate when the environmental conditions for learning and recall were the same. Supports contextual cues for the encoding specificity principle.

Loftus and Palmer 1974 - leading Qs

45 American students formed a volunteer sample and took part in a lab study where they watched videos of traffic accidents and were asked questions regarding the film. One of the questions was leading and the IV was the different verbs, “How fast were the cars going when they [smashed/hit/collided/contacted]?” Smashed = 40.8mph. Contacted = 31.8mph. Ps were also asked 1 week later if they saw any broken glass, Ps who were asked smashed were more likely to report seeing broken glass. Shows leading questions can influence your memory of an event.

Gabbert et al. 2003 - post event discussion

Participants watched a video of a girl stealing money from a wallet. The co-witness group (2 Ps) discussed the event together before completing a questionnaire. Unknown to them, they had seen the same crime from different perspectives, so only 1 person had actually witnessed the girl stealing. Results showed 71% recalled details they not seen from their perspective, and 60% identified the girl as guilty despite not having seen her commit the crime. Highlighting the impact of post event discussion on memory accuracy/EWT.

Johnson and Scott 1976 - Weapon focus

Ps were sat in a waiting room outside a lab waiting for the “real” study to start. There were two conditions: Ps heard a discussion in the lab followed by a man leaving holding a pen covered in grease; Ps heard an argument followed by a man leaving holding a knife covered in blood. The Ps were later asked to identify the man from a photo line up. 49% correctly identified the man holding the pen, but only 33% could identify the man with the bloodstained knife. They argued that Ps showed worse recall for the man with the knife because their attention was fixed on the knife; weapon focus.

Christianson and Hubinette 1993 - anxiety EWT

Natural study. Interviewed witnesses to a Swedish bank robbery months after the crime. Recall of the events was more than 75% accurate for all witnesses. But victims who were directly involved, and experienced higher levels of anxiety showed a higher accuracy than bystanders. This suggests anxiety increases accuracy of recall.

Patient EVF - opposes CE

EVF suffered brain damage, could reason but not make decisions, suggesting the central executive is more complicated than previously thought, alluding to subcomponents. Limitation of WMM.

Tulving and Psotka 1971 - Forgetting

Found retrieval failure is a better explanation of forgetting than interference. Because when given multiple word lists and asked to remember the first one, recall was worse the more lists they had viewed since the first. However, when given cues, recall was 70% regardless of how many lists they had been given. Suggesting retrieval failure is a better explanation for forgetting.

Goodwin et al. 1969 - state cues

Taught male participants a set of words, either drunk or sober. They were asked to recall the words 24 hours later, when either drunk or sober again. He found participants showed better recall when in the same state as they learned the words.

Kohnken 1999 - Cognitive interview

Kohnken conducted a meta-analysis comparing the cognitive interview to a standard interview. He found an increase in both correct (81%) and incorrect (61%) information for the cognitive interview.