Cognitive Psychology - Memory

Key study: Peterson and Peterson

How

• The participants were 24 students. Each was tested over 8 trials.
• On each trial a participant was given a consonant syllable and a three digit number eg. THX 512.
• They were asked to recall the consonant syllable after a retention interval of 3,6,9,12,15 or 18 secs.
• During the retention interval they had to count backwards from their three digit number.

Showed

• Participants on average were 90% correct over 3 seconds.
• They were 20% correct after 9 seconds.
• They were 2% correct after 18 seconds.

Suggesting that STM has a very short duration, less than 18 seconds, if verbal rehearsal is prevented.

Peterson and Peterson (Point, Elaboration, Link)

• 1) Study lacked ecological validity because the stimulus material was artificial.
• Trying to memorise consonant syllables does not truly reflect most everyday memory activities where what we are trying to remember is meaningful.
• However, we do sometimes try to remember fairly meaningless things, eg. pin numbers and post codes.
• So study does have some relevance to everyday life.

_

• 2) The findings may be explained by displacement rather than the material only lasting a short time in STM.
• Participants were counting the numbers in their STM and this displaced the syllables to be remembered.
• Reitman used auditory tones instead of numbers so that displacement wouldn't occur She found that duration of STM was much longer than P&P's study.
• So study was likely to be due to displacement rather than decay.

Key Study: Bahrick et al.

How

Tested 400 people of various ages (17-74) on their memory of classmates. There were various tests, including:

• Photo recognition test consisting of 50 photos from the participants high school yearbook.
• A free-recall test where participants were asked to list all the names they could remember of individuals from their high school.

Showed

Participants who were tested within 15 years of graduation were about 90% accurate in identifying faces.

After 48 years, this declined to about 70% for photo recognition.

Free recall was about 60% accurate after 15 years, dropping to 30% after 48 years.

Bahrick et al. (Point, Elaboration, Link)

• 1) Results may be due to rehearsal.
• For example, some participants might have looked at the yearbooks regularly and that's why their recognition/recall was good.
• Rehearsal is acting as an extreanous variable.
• This means that Bahrick et al's findings lack validity.

Nairne et al.

How

• Nairne et al. modified the Peterson's technique so that the material to be remembered was the same across trials. This was done to prevent one set of items interfering with another, which might affect recall.
• Participants were shown five nouns and, after a retention interval, were shown all five nouns in a different order and had to recall the nouns in the correct order.

Showed

• Found that items could still be accurately recalled after as long as 96 seconds.
• Therefore it seems that information remains for longer in STM if there is no interference from other items.

Evaluation

• One criticism is that words were used instead of consonant syllables.
• Participants might be better at remembering real words than meaningless consonants or numbers.
• This means that these results may only apply to some kinds of STM tasks.
• However, the results still show that forgetting rates in STM vary depending on the presence or absence of interfering materials from prior trials.

Key Study: Miller

How

Miller made observations that everyday things come in sevens eg. seven notes on the musical scale, seven days of the week.

Miller also reviewed several studies that have investigated the span of STM, eg. where participants counted dots flashed on a screen or were tested on the recall of words.

Showed

• When dots are flashed on a screen, participants are reasonably accurate when there are seven dots but very inaccurate when 15 dots are shown.
• This suggests that the span (or capacity) of STM is about seven items (plus or minus 2).
• Miller also observed that people can recall five words as well as they can recall five letters. They do this by chunking - grouping sets of digits or letters into meaningful units.

Miller (Point, Elaboration, Link)

• Miller may have overestimated the capacity of STM.
• Cowan reviewed research on this topic, and found that the capacity of STM is more likely to be about four chunks rather than seven chunks.
• Cowan's finding was further supported by Vogel et al, who found four items was about the limit for visual items.
• This means that the lower end of Miller's range is more appropriate (7-2 which is 5).

_

• + There are real-world applications to the UK postcode system.
• Baddeley used research on chunking to make recommendations about what combinations of letters and numbers are most easily rememebered.
• Baddeley found that numbers were best remembered if placed between the city code and some random letters. Phone numbers and car license plates also use chunks.
• This shows that research on chunking has led to useful applications.

Key Study: Jacobs

How

Jacobs tested digit span.

Each participant first listened to four digits and had to recall these in the correct order. If this was correct they progressed to five digits and so on, until they couldn't recall the order correctly. This determined the participant's digit span.

The same activity was repeated with letters.

Showed

• Jacobs found that the mean span for digits was 9.3 items whereas it was 7.3 for letters.
• He also observed that digit span increased with age; 8 year olds could remember an average of 6.6 digits whereas the mean for 19 year olds was 8.6 digits.
• This might happen because people develop strategies to improve their digit span as they get older, such as chunking.

Jacobs (Point, Elaboration, Link)

• 1) Jacobs study is very old because it was conducted in 1887.
• Jacobs may not have adequately controlled the study because research was not as formal.
• The instructions he used with adults and children might have been differed, acting as an extraneous variable.
• This might explain why adults performed better, rather than there being any real difference.

Simon

How

Simon tested his own recall for one-syllable, two-syllable and three-syllable nouns and also familiar phrases. He read them aloud and then later tried to see how many he could recall.

Showed

• He found there was a slightly shorter span for three-syllable words than one- or two- syllable words.
• He also found that there was a shorter memory span for longer phrases than shorter phrases.

Evaluation

• Study may lack generalisability because only one person was involved.
• However, in this particular situation it may not matter that only one person's data was observed as the findings fit with common experience.
• The data is more of an observation than a formal research study.
• However, it refines our understanding of chunking.

Key study: Baddeley

How

• Group A participants were given acoustically similar words (eg. cab can mad max).
• Group B were given acoustically dissimilar words (eg. pit, few, cow, pen).
• Group C were given semantically similar words (eg. great, large, big, huge).
• Group D were given semantically dissimilar words (eg. good, huge, hot).

On each trial 5 words were read out. After a time interval, participants were shown 10 words and asked to select the correct 5 words and place them in the right order.

Showed

• When STM was tested, participants with acoustically similar words had lowest recall.
• This suggests that words in STM are encoded acoustically.
• When LTM was tested (time interval of 20 minutes), the group with semantically similar words had the lowest recall.
• This suggests that words in LTM are encoded semantically.

Baddeley (Point, Elaboration, Link)

• 1) Artificial stimuli were used.
• This means we should be cautious about generalising the findings to different kinds of memory task.
• For example, if people were processing meaningful information they might use semantic encoding even for STM tasks.
• Meaning that this study only tells us a limited amount about encoding in STM/LTM.

_

• 2)The LTM memory task wasn't very long term.
• It was only 20 minutes, which is not the same as remembering information for months and years.
• It is possible that different kinds of encoding processes are involved when information is stored over months and years.
• However, this study does clearly show there is a difference over time.

Key study: Brandimote et al.

How

• Brandimote et al. showed participants six picture pairs.
• One of each pair was a component of the other.
• Participants were asked to 'subtract' the second picture from the first one for each pair.
• In the retention interval some participants had to say 'la la la la' to prevent any verbal rehearsal (an articulatory suppression task).

Showed

Participants performance was unaffected by the articulatory suppression task.

This shows that the images were visually rather than verbally encoded and that encoding in STM is not always acoustic.

Normally we 'translate' visual images into verbal codes in STM. However, since verbal rehearsal was prevented in this study, visual encoding was used.

Brandimote et al (Point, Elaboration, Link)

• 1) + Brandimote et al.'s study was supported by other studies
• Wickens et al. found that STM sometimes uses a semantic code.
• Frost's study showed that LTM was related to visual as well as semantic categories.
• Therefore encoding is not simply acoustic or semantic but can vary according to circumstances.