Encoding process (according to Atkinson and Shiffrin's multi-store model of memory)
Information from the senses must be encoded to enable transfer from the initial SensoryRegister to ShortTerm Memory (STM), and in turn onto LongTermMemory (LTM)
Baddeley (1966) found that more mistakes are made when recalling acousticallysimilar words straight after learning them, whilst more mistakes are made when recalling semanticallysimilar words 20 minutes after learning them (LTM recall)
Miller's idea that things come in groups of 7 suggests that we are predisposed to remembering this quantity and that such a 'chunking' method can help us recall information
Jacobs demonstrated that the mean letter span was 7.3 and the mean digit span was 9.3 (i.e. the number of letters or digits we can recall after increasing intervals)
Petersen et al (1959) found that increasing retention intervals decreased the accuracy of recall of consonant syllables in 24 undergraduates, when counting down from a 3 digit number (preventing mentalrehearsal)
Bahrick et al (1975) found that photo recognition of graduating classmates of the 396 participants decreased from 90% to 70% between 15 years and 46 years of graduating
Current laboratory experiment methodology produces highly reliable and valid data through controlling and so removing the effects of extraneous and confounding variables
Bahrick et al's findings have high ecological validity because they can be generalised to real-life, due to the stimuli reflecting those which we would often try to learn and recall in day to day lives: information with personal and meaningful value
Petersen et al and Miller et al studies feature methodologies with low mundane realism, thus producing findings with low ecological validity due to use of artificial stimuli which has low personal meaning to the participants
More recent research has suggested that Miller may have over-exaggerated the capacity of short-term memory, and that the capacity is more similar to 4 chunks as opposed to the original 5-9 limit
Miller's outdated methodologies and lack of control over confoundingvariables may have contributed to this inaccurate estimate of short-term memory capacity