Forgetting occurs in the STM as it has a limited duration and capacity; once these limits are reached, information is forgotten. If information is forgotten from STM therefore it is unavailable, however LTMs duration and capacity are theoretically unlimited so any information that is forgotten from here is only inaccessible in memory somewhere but not retrievable for some reason.
Decay from STM is where chemical memory traces fade after 15-30 seconds without rehearsal, so presumably some sort of structural change takes place during learning, and according to decay theory, metabolic processes occur over time which cause the engram (memory trace) to degrade unless maintained by rehearsal, causing the memory to become unavailable.
Peterson and Peterson (1959) conducted a study which supports decay theory as the average number of trigrams recalled was high when there was a short delay in recall, and nearly 70% of trigrams had been forgotten after a 9 second delay suggesting the duration of STM is only a few seconds and that decay can take place due to forgetting things over time. However, there are difficulties with this study and decay theory in general as other effects need to be excluded. Ideally, a study should get participants to receive information then do nothing physically or mentally for a set amount of time and then test recall, but this is impossible.
Jenkins and Dallenbach (1924) did a study which approximated the do nothing state by asking participants to learn a list of syllables before they either went to sleep or continued with their normal activities. Recall was tested at regular intervals, and it was found that forgetting increased as retention intervals increased for awake participants but not for the sleeping participants, but if decay is from time passage alone, there should be the same amount of forgetting in both groups, suggesting what occurs between learning and recall determines orgetting, not time passage.
Some data exists suggesting neurological breakdown occurs with age and disease (e. g. Alzheimers) however there is no evidence that the major cause of forgetting in LTM is neurological decay (Solso 1995). Hebb argued that while learning is taking place the engram that will eventually be formed is delicate and liable to disruption (the active trace). With learning it grows stronger until a permanent engram is formed (structural trace) through neurochemical and neuroanatomical changes.
The active trace corresponds roughly to STM and according to decay theory, forgetting from STM is due to disruption of the active trace. Other researchers argued it can explain LTM forgetting if it is assumed decay occurs through disuse (decay-through-disuse theory), so if certain knowledge or skills arent used for long periods of time the corresponding engram will eventually fade away (Loftus and Loftus 1980). Reitman (1974) argues that items are displaced from the STM.
New information coming in pushes old information out of the limited capacity STM (7 +/- 2 items Millers Magic Number) before it has been rehearsed and transferred into LTM. Waugh and Normans 1965 study supports this theory. They used a serial probe technique where participants were presented with 16 digits, then a probe digit was read out, and the participants had to say which digit came after the probe digit in the list. They found that 11 Spicy Strawberry information at the start of the list was forgotten more than the information at the end, presumably because information at the start had been displaced by newer information.
It is assumed that if the probe was towards the end of the list the probability of recall was high, as the last digits would still be available in the STM. When the number of digits following the probe was small, recall was good but when it was large recall was poor, consistent with displacement theory. The primacy-recency effect also provides evidence for displacement. Glanzer and Cunitz conducted a study to demonstrate this effect where if recall is delayed the effect of recall disappears.
If asked to remember a list of words, words from the beginning and end of the list tend to be remembered better than words in the middle. The high level of recall at the beginning is the primacy effect, where your short term memory is quite empty giving room to rehearse and pass information to the long term memory store. However as words keep coming the short term memory store loses ability to rehearse information because it becomes overloaded, meaning information at the end of the list is also recalled and not words in the middle of the list due to the new words displacing old words.
However, it is possible that the information at the start decayed with time and was not displaced so it is not clear which is the more likely explanation for forgetting in the STM. Forgetting from LTM is suggested to occur from retrieval failure, and the Encoding Specificity Principle states that when information is learned, other information such as place of learning is encoded at the same time. Where external and internal cues for remembering are lacking such as context or state, forgetting occurs.
Context-dependent forgetting occurs when the environment is different to where information was originally learned, and state-dependent forgetting is where your mood is not the same as when information was learned, and it is relevant as an internal cue (McCormick and Mayer). The role of retrieval cues is demonstrated by the tip of the tongue phenomenon, where we know something but cant retrieve it at that moment in time. Brown and McNeill (1966) investigated this, and gave participants dictionary definitions of words and asked them to give the word they were describing.
Some were sure they knew the word but couldnt recall it, suggesting the required words were in memory but an absence of a correct retrieval cue prevented recall. Godden and Baddeley did a study where divers learned a list of words either on land or underwater and were later tested for recall either on land or underwater. They found that divers who learned words in the same environment they recalled them performed better than those who recalled words in a different environment, which suggests that recall of information is better when in the same context of where it was learned.
However, Godden and Baddeley repeated their study using recognition as a measure of remembering but found that context had no effect, which may mean that context affects recall only. Tulving and Pearlstone showed that cued recall is more effective than free recall. Category names with words inside where given as cues in one condition, and for the other condition words were asked to be recalled without category names as a cue. 0% of words were remembered accurately from free recall, and 60% were accurately remembered from cued recall, showing that cues do help memory.
This supports retrieval failure theory as it offers a method of remembering information more effectively if there is something present to trigger it. Miles and Hardman also support this theory as participants learned words either at rest or while exercising, and recall was tested in the same or different state to learning and it was found that physical state provides cues to assist recall.
Bower et al supports McCormick and Mayers theory of mood as an internal cue to recall, as participants were hypnotised and imagined a happy/unhappy mood while learning information and it was found that 12 Spicy Strawberry participants who recalled the material in the same mood as they learnt it performed better than those in a different mood, so recall was affected by internal context in which they learned the material.
Although experiments conducted in a laboratory have low ecological validity due to material required to be recalled and the controlled environment, Eysenck says it has been proved easy to demonstrate cue-dependent forgetting outside the laboratory. Interference theory in LTM is another explanation for why we forget. There are two types retroactive interference where new material disrupts the recall of old material and proactive interference where old material disrupts the recall of new material.
The more similar the material, the greater the interference. This may be why it is generally agreed that if students need to study more than one subject in the same time frame they should be as dissimilar as possible to prevent interference from occurring. Baddeley and Hitchs 1977 study supports the idea of interference as an explanation for forgetting in the LTM. They asked rugby players to recall the names of teams played earlier in the season, and some players played less games than others in the intervening period due to injuries.
Players memory of two weeks ago was worse when they played more games compared with players who didnt take part in as many games. Since the same time had elapsed for all players, forgetting seemed to be due to interference of new learning (retroactive interference), showing that new memories can disrupt old memories of what happened and cause confusion and inaccuracies and that forgetting is influenced more by what we do before or after learning than by passage of time.
Keppel and Underwood said that proactive interference was occurring in the Petersons study into decay theory as the ability to remember trigrams had been disrupted by the old information as trigrams may interfere with each other, so it may not be due to time passing. Wickens found that participants became increasingly poor at retaining information in STM on successive trials, however when the category was changed, performance was as good as for the first list. So performance with lists of numbers became poorer over trials, but if the task was changed to lists of letters it improved (release from proactive interference).
Interference theory is very relevant to us as there are many real-life examples of where it can affect memory, e. g. not remembering new phone numbers due to having an old number for so long, or calling a current partner by your old partners name by accident. These are both examples of proactive interference. However, as with most laboratory experiments, they are low in ecological validity due to the controlled environment and learning being artificially compressed in time which maximises the likelihood that interference will occur.
Laboratory studies also tend to use nonsense syllables as a stimulus, so when meaningful material is used, interference is more difficult to demonstrate (Solso 1995). According to Schacter (2002), efficient forgetting is crucial to a fully functioning memory. In his book The Seven Sins of Memory, he describes several ways we forget, including transience discarding out of date information, absent-mindedness failing to properly encode information due to a lack of attention, and blocking where the brain holds back on a memory in favour of a competing memory, preventing interference.