AQA Psychology A Unit 1 - Key Studies

Concluded on a capacity of 12 for the sensory memory after participants could recall an average of three letters from any row of a 4x3 grid of letters, suggesting all were available in the sensory memory (the others couldn't be recalled afterwards because they had decayed due to short duration).

Conducted the consonant trigram retention experiment, finding a duration of 18 seconds or less for the short-term memory store. Participants were given three unpronounceable consonants then asked to repeat the trigram after counting backwards for a gradually longer time.

We can hold (capacity) 7±2 bits of information in the short-term memory.

Put forward the multi-store model of memory, with sensory, short-term and long-term memory.

An amnesiac who provides support for the multi-store model - memory being unitary explains why he can still walk and play the piano but cannot lay down any new memories.

Introduced the working memory model, including sensory memory, the working memory (central executive, phonological loop, visuo-spatial scratch pad, episodic buffer) and long term memory.

In a real bank robbery (high ecological validity!) those who were threatened had a more accurate recall of the event, showing that higher anxiety leads to greater accuracy of recall.

The weapon focus effect. From 50 photographs, 49% accuracy in identifying man when he had pen and greasy hands, 33% with knife and bloody hands.

Over 600 adults stopped in street and asked to describe appearance of woman spoken to 2 minutes earlier. Younger participants more confident, but no difference in accuracy found between younger and older people.

With a delay of 35 minutes, there was no difference in accuracy. With a delay of one week, the older group was significantly less accurate in its recall.

Participants of all ages asked to rate pictures for attractiveness. Then asked to identify those in pictures they'd just rated. Older people less accurate. All groups most accurate in identifying individuals of their own age group (own-age bias).

45 students watched a video of a car crash and asked to estimate speed of cars involved. When verb in question was 'smashed', mean estimate over 40mph. With verb 'contacted', mean estimate just over 30mph. Shows the effect of leading questions/post-event information.

When verb was 'smashed', many more said, a week later, they'd seen broken glass than when verb was 'hit' (there was no broken glass in the video).

Developed the cognitive interview, which includes context reinstatement, report everything, report in reverse order and report from another perspective.

In a regular interview, participants are bombarded with questions that don't allow them to speak and do not accord with the individual's memory of the event.

Found less inaccurate items, more accurate items but more confabulated items produced by a cognitive interview, in comparison with a regular interview.

Participants tracked a moving light whilst performing an imagery task (analysing letter F corners) - couldn't do both accurately simultaneously. Could do light tracking with a verbal task. Concluded first two tasks both tried to use visuo-spatial scratch pad.

Research support for external, context dependent retrieval cues. Participants given list of 80 words in a basement room. Following day, average recall was higher for participants back in basement room than for those out of it.

Participants read 5 words, either acoustically or semantically similar or dissimilar, and asked to repeat them in same order. Higher accuracy with acoustically dissimilar words (due to acoustic confusion with similar ones). A small negative effect with semantically similar words. Mainly acoustic encoding in STM.

Participants given a visual task of subtracting from an image and identifying the new image. Participants were actually more accurate when also repeating a chant (which stopped them acoustically encoding the images). Sometimes, then, visual encoding is used in STM.

Argued for a 'permastore' (unlimited duration of LTM) through findings of around 60% accuracy in identifying pictures from high school yearbook almost 50 years after graduation. Argued this showed some information may be kept for up to a lifetime.

Participants read a list of 10 words, either semantically or acoustically similar/dissimilar. After 20 minutes and completion of a separate task, recall was far better for semantically dissimilar words than for similar ones (with acoustic differences having little effect). So, LTM encoding mainly semantic.

150 participants watched a film of a car accident and were then split into two groups. Group 1 was asked to estimate the speed of the car when passing a 'stop' sign. Group 2 asked the same for when car was passing a barn (not in video). A week later, about 2% of Group 1 said they'd seen a barn in the video, compared with around 17% of Group 2.

Conducted a meta-analysis of 27 studies on the cognitive interview schedule. In all studies, it provided more accurate information and recall than standard interviews did.