MEMORY

Geiselman 1984

1) Mental reinstatement of original context -  The interviewer encourages the interviewee to mentally recreate both the physical and psychological environment of the original incident. The aim is to make memories accessible, since people need appropriate contextual and emotional cues to retrieve memories.                                                                                                    2) Report everything - The interviewe encourages reporting every single detail of the event without eiditing anything out even though it may seem irrelevant. Memories are intercollected with one another so that recollection of one item may then cue a whole lot of other memories. Also small details could then be pieced together to form a clearer picture.                                            3) Change order - Maybe reversing the order in which event occured. This is because our recollections are influenced by schemas e.g. if you went to a restaurant a few weeks ago your recollection will be influenced by your general expectations of what is likely to happen at a restaurant. If you have to recall event starting from the end of the event backwards, this prevents your pre-existing schema influecing what you recall.                                                                   4) Change perspective - Asked to imagine how it wouldve appeared to other witnesses present at the time. This again is done to disrupt the effect that schemas have on recall. Anderson and Pichert 2009

Effectiveness - A meta analysis of 53 studies found on averaage an increase of 34% of the amount of correct information generated in the CI compared with the standard interview Kohnken 1999, although most of these studies involved volunteer witnesses (usually college students) tested in a lab (such studies do not reflectreal world practices). Effectiveness may be due to some individual elements ratther than the whole thing. Milne and Bull 2002 interviewed under gard students and children using just one individual component of the CI nand compared the responses gathered to a control condition (where they were simply asked to try again) Recall was boradly similar and  no different to the control group. When participants were interviwed using aa combination of the report everything and mental reinstatment, their recall was significantly higher .                                                                                                                                              Quantity vs quality - The procedure is designed to enhance the quantity of correct recall without comprimising the quality (the amount of correct recall as a % of total recall) of that information. However it may be that effictiveness has largely been in terms of quantity. Kohnken found an 81% increase in correct info but also a 61% increase of incorrect info when CI was compared to the standard interview. This means all the info needs to be treated with caution since it does not garuntee accuracy.                                                                                                                                             

  Problems with using the CI in practice - Kebbel and Wagstaff said police officers said this technique requires more time than is often available and  that instead they prefer to used stratergies to limit and eyewitnesses report to the minimum amount of infor they feel is necessary. Also CI requires special training and many forces have not been able to provide more than a few hours.                                                                                                                                      

Individual differences - CI may be useful when interviewing older witnesses. Negative stereotyppes about older aadult memories can make witnesses overly catious about reporting info. However with the CI since it stresses the importance of reporting everything may overcome such difficulties. Mello and Fisher compared older 72 years and younger 22 years adults memory of a filmed crime using CI or standard interview. The advanatge of CI was  greater for the older than the young.

Key study : Johnson and Scott 1976 Procdure - They asked participants to sit in a waiting room where they heard this argument in an adjoining room and then saw a man run through the room carrying either a  pen covered in grease (low anxiety) or a knife covered in blood  (high anxiety). Partcipants were lateer asked to identify the man.

Findings - The findings support the idea of weapon focus effect. Mean accuracy was 49% identifying the man with the pen and 33% with the knife. Loftus 1987 showed that anxiety does focus attention on central features of a crime (weapon). The researcher monitored peoples eye movements and found presence of a weapon cause attention to be phsycially drawn towards the weapon itself and way from the persons face.

Anxiety has a postive effect on accuracy  - Christianson and Hubinette 1993 found ehnanced recall when they questioned 58 real witnesses to a bank robberies in Sweden. The witnesses were either  victimes(bankteller) or bystanders (customers) high and low anxiety. The interviews were conducted 4-15 months after the robberies. All witnesses showed good recall 75%. Those victims who were most anxious has th best recall of all.

Resolving the contradiction  - 

Weapon focus may not be caused by anxiety - Pickel siad  weapon focus effect could be due to suprise rather than anxiety. She arranged for partcipants to watch a theif enter a hairdresser salon carrying scissors (high threat,low surprise) handgun(high threat high suprise) wallet (low threat low suprise) or a whole raw chicken (low threat high suprise). Identification was least accurate in the high suprise conditions rather than high threat.                                                 Real life vs lab studies - A strength of Christianson and Hubinette is that the crime was real. It may well be that lab studies do not create the real levels of anxiety experienced by a real eye witness during an actual crime. Deffenbacher agreed with this but  found from a review of 34 studies , that lab studies in general demostrate that anxiety leads to reduced accuracy and that real life studies lead to an even greater loss in accuracy, which is at odds with the result of C&H.                                                                                                                                                                                     No simple conclusion - Many studies of anxiety did not involve  violence, like C&H. Halford and Miller found that victims of violent crimes were more accurate in their  recall of crime scene in formation than victims of non violent crimes. This shows their is no simple rule about the effects of anxiety on the  accuracy of EWT.                                              Individual differences - One exttraneous variable is emotional sensitivity. Bothwell tested partcipants for neuroticism a personality characteristic where individuals tend to become anxious quickly. They were labelled as Neurotic or stable (less emotionally sensitive).Found that stable participants showed rising levels of accuracyas stress levels increased whereas the opposite was true of neurotics.                                                                                                                              Demand characteristics - Most lab studies show participants a video of a filmeed and usually staged crime. These partcipants would then work out that they will be asked about what they have seen so they will pay alot more attention.

Yerkes-Dodson effect curve said when anxiety is too high or too low then performance will be decrease but if anxiety is at a moderate level performance will increase and be better.

Leading  question : Loftus and Palmer 1974. Experiment 1 Procedure - 45 students were shown 7 films of different care accidents. After each film the participants weregiven a questionaiire which included 1 criticaal questions 'How fast were the cars going when they hit eachother?' One group of partciants were giivent this question and the other 4 were given the verbs 'smashed, collided,bumped,contacted'.

Findings - Smashed 40.8, collided 39.3, bumped 38.1, hit, 34, contacted 31.8.

Experiment 2 Procedure - The leading question may bias particippants response or may actually cause info to be altered before it is stored. A new set of partcipants were divided into 3 groups and shown a film of a car accident lasting 1 min. The particpants returned 1week later when they were asked 10 question, 1 critical question was 'Did you see any broken glass?' There was no broken glass but those who thought the car wastravelling faster would be more likely to think there  was.

Post event discussion - Conformity effect - Gabbert 2003. Participants were in pairs where each partner watched a different video of the same event so they each viewed unique items. Pairs in one condtion were encouraged to discuss the event before each partner individually the event they watched. 71% of participants who discuessed the event went on to mistakenly recall items acquired during the discussion.  

Supporting evidence - Loftus, college students were asked to evaluate advertising material about Disneyland. Embedded in this material was misleading info about Bugs bunny or Ariel. Participants were assigned to the Bugs, Ariel or a control condition (no misleading info). All had visited disneyland, participants in the Bugs or Ariel group were more likely to report they shook hands with these characters, so misleading info can create false memory.                            EWT in real life - Lab experiments such as those carried out by Loftus may not represent real life because people dont take the experiment seriously and/or they are not emotionally aroused in the way they would be in a real incident. Foster 1994 found that if participants thought they were watching a real life robbery and also thought their responses would influence the trial, the identication was more accurate. Yullie and Cutshall also found greater accuracy in real life. Witnesses to an armed robbery in Canada gave very accurate reports of the crime 4 months after the event even though they were given 2 misleading questions. Misleading info has less influence in real life EWT.                                 Real world application - The criminal justice system relies heavily on eyewitness identification. DNA shows that mistaken eyewitness identification was the largest single factor contributing to the conviction of innocent people Wells and Olsen.                                                                                                                                                                Individual differences - There is evidence that older people are less accurate than younger people when giving eyewitness reports. Older people find it difficult to remember the source of this info, even though their memory of the info itself isnt impaired, as a result they are more prone to misleading questions since they are less sure. Schacter.        Demand charcteristics - Zaragose and McCloskey said that many answers participants give in lab studies are because of demand characteristics. Participants dont want to let the researcher down and want to appear helpful and attentive, so when asked a question they dont know the answer to, they guess ecspecially if its a yes or no question.

The encoding specificity principle - Tulving and Thompson  1973 proposed that memory is most effective if info that was present at encoding is also available at the time of reteval. The cue doesnt have to be exactly right but the closer it is to the original  item the more useful it is. Tulving and Pearlstone 1966, participants had to  learn 48 words belonging to 12 catergories. Each word was presented as catergory+word e.g. fruit-apple, parrtcipants either had to recall as many words as they could (free recall) or they were given cues in the form of catergory names (cued recall). 40% in free recall and 60% in cued recall. There is another type of cue which is not relatng to the learning material in any meaningful way. When onfo is learnt we also remember where we were (environmental context). How we felt.

Context-dependent forgetting - Abernethy 1940 , she arranged for a group of students to be tested before a course began and they were then tested each  week of their course. Some students were tested in their teaching room with their usual instructor whereas others by different instructors, or in a different room. There were 4 conditions. Those who were tested in the same room by the same nstructor performed the best. Godden and Baddeley 1975 testing contextual cues, participants had to learn words either on land or underwater. They were then tested on land or underwater, so 4 condtions. Highest recall when initial context matched the recall  environment.

State dependent forgettng - Goodwin asked male volunteers to remember a list of words when they were drunk or sober. They were asked to recall the words after 24 hours when they were sober or drunk again. Info learnt when drunk is more avalable in the same  state later.

There is alot of research support - There is alot of research evidence on the facing page. Such research includes lab, field and natural experiments and anecdotal evdence so it has relevance to everyday memory experiences. Increases validity. 

Real world application - Use this info to improve recall when you need it e.g. in exams. Abernerthy said you should revise in the room you are taking the exam in. You could use imagination to acheive this. Smith said just thinking of the room where you did the original learning was as effective as actually being n the same room. Retrieval cues in cognitive interview.

Retrieval cues do not always work - Is not that effective in exams, because the info you are learning is related to alot more than just cues, because when we learrn we learn about stuff such as Milgrams study or MSM which are more complex than word lists and are less easly trggered by cues. 

Retrieval failure explain interference effects - Tulving and Psotka , participants were given 6 different word lists to learn each consisting of 24 words divided into 6 catergories (type of tree,flower). After each list was presented the participants were asked to write down as many word as they could remember (free recall). After all the lists were presented they were gven catergory names and again asked to recall words from all the lists (cued recall). Some participants only learnt one list and the others 2 and so on. The more lists the participants learnt the worst their performance would become (retroactive interference). However when participants were given cued recall the effects of interference had disappeared - remembered about 70% of words. This shows info is availablebut cannot be retrieved and that retrieval failure is more a more important explanation of forgetting than interference.

Retroactive interference - Muller and pilzecker 1900 gave paricipants a list of nonsense syllables to learn for 6 minutes and then, after a retention interval, asked partcipants to recall the lists. Performance was less good if participants had been given an intervening task between initial learning and recall (they were shown 3 landscape painting and asked to describe them. The intervening task produced RI because the later task interferred with what had previously been learned.                                                                                                                                           Proactive interference - Underwood 1957 analysed the findings from a number of studies and concluded that when participants have to learn a series of word lists they do no learn the lists of the words encountered later on in the series as well as lists of words encountered earlier on.  Underwood found that participants memorised 10 or more lists, then after 24 hours they remembered about 20% of what they had learnt. If they learnt one list then recall was 70%.

Similarity of test materials - McGeoch and McDonald 1931 gave particpants a list of 10 adjectives (list A). Once these were learnt there was a resting interval of 10 minutes during which they learned list B, followed by recall. If list B was synonyms of list A, recall was poor 12%, if list B was nonsense syllabus this had less effect 26% recall, if list B was number it had 37%. This shows interference is the strongest the more similar the items. Only interference rather than decay can explain such effects.

A real world study  - Baddeley and Hitch 1977, investigated rugby players recalling the names of players they had played against over a rugby season. Some players played in all the games but some missed games because of injuries. If interference theory is correct then those players who played most games should forget proportionately more because of interference - which is what Baddeley and Hitch found.

Research is quite artificial - Most research is lab based and often use artificial lists of word or nonsense syllables. Thus the findings may not relate to everyday uses of memory. Also participants may lack motivation to remember links in such studies and this may allow interference effects to appear stronger than they really are - low ecological validity. The counterargument is interference effects have been obserevd in everyday situations as shown on other page.                                                                                                                                                         Interference only explains some situation of forgetting  - Interference effects dont occur that often in everyday life. Rather special conditons are required for interference to lead to forgetting - the 2 memories need to be similar. So interference is considered to be unimportant for explaining forgetting. Anderson 2000 said how much forgetting can be attributed to interference remains unclear.                                                                                        Accessibility vs availability - Ceraso 1967 found that, if memory was tested again after 24 hours, recognition (accessibility) showed considerable spontaneous recovery, whereas recall (availibility) remained the same. This suggests that interference occurs because memories are temporarily not accessible rather than actually having been lost.                                                                                                                                                                        Real world application to advertising - Danaher 2008 found that borh recall and rocognition of an advertisers message were impaired when participants were exposed to 2 advertisements for competing brand within a week. This is a serious problem considering the amount of money spent on adverts. Danaher suggests that running multiple exposures to an advert in one day rather than spread these over a week. This will result in reduced interference from competitors adverts.                                                                                                            Individual differences - Kane and Eagle 2000 demonstrated that individuals with a greater working memory span were less suspectible to proactive interference. They tested this by giving participants 3 word lists to learn. Those participants with low working memory spans showed greater proactive interference when recalling the second and third word lists than did people with higher spans.

Episodic memories - This is your own personal experiences or episodes in your life that is personal to you e.g. your first day at school, your first kiss etc. You may recall the time, place, who was there or the context such as what happened just before or after or emotions you felt at that time. so 3 element: specific details of the event, the context and the emotion.

Semantic memory - Knowledge about the world that is shared by everyone, may relate to social customs, or abstract concepts such as maths and language. Semantic memories generally begin as episodic memories because we acquire knowldge based on personal experiences. The transition happens when the memory slowly loses it association to particular events, so that the info can be generalised as semantic memory. Sometimes people do have a strong recollection about when and where they learned a paricular fact.

Procedural memory - Its skills such as tieing shoelaces or playing the piano, or knowing how to read. They are typically acquired through reptition and practice. This memory is automatic -  so if you think about the memories when carrying them out you, it prevent you from acting it out.

Evidence of brain scans - Different areas of the brain are active when the different kinds of LTMs are active. Episodic memory is associated with the hippocampus and other parts of the temporal lobe and the frontal lobe. Semantic memory relies on the temporal lobe. Procedural is Cerebellum. Tulving 1994 found left prefrontal cortex was invlved in recalling semantic memories and right prefrontal cortex cortex in episodic memory. This has been confirmed in othe research studies confirming the validity of the studies.

Clinical evidence - HM and Clive wearing had brain damage and epsidoic memory in both men was severly impaired. Their semantic memory was relatively unaffacted e.g. they still understood the meanings of words. Both had procedural memories intact Wearing still knew how to play the piano. Supports Tulvings view that there are different areas of the brain for each type of memory so  one store can be damaged but the others are unaffected.However brain injury is traumatic and patients may have difficulty paying attention and underperform on taks. Also you cannot be sure of the exact part of the brain affected until the patient has died because that damage to a particular area of the brain doesnt mean that the area is responsible for a particular behaviour - it may just be acting as a relay station. Also wont know if the patients memory was bad before the accident.

Distinguishing between episodic and semantic memories - Hodges and Patterson studied patients with alzheimers and found some patients who have the ability to form new episodic memories but not semantic - This is a single dissociation i.e. a speration between 2 abilities. But episodic memories place greater demands on mental processing and thats why its more effected by brain damage. So researchers look for double dissociation, Irish 2011 found that some patients had poor semantic memory and intact episodic memory. This suggests that sematic memories can form seperately.

Baddeley and Hitch 1974 felt that STM was not 1 store but a number of different stores because if you do 2 things at the same time (dual task performance) and they are both visual tasks, you perform them less well than if you do them seperately. If you do 2 tasks and one is visual and one is sound, then you do them equally as well. This suggests there is one store for visual processing and one for sound.

Central executive - Its function is to direct attention to particular taks, allocates tasks to other slave systems, data arrives from the senses or from LTM. The CE has a very limited capacity.                                                                            Phonological loop - Also has a limited capacity, it deals with auditory info and preserves the order of the info. Baddeley 1986 further subdivided this loop into 1) The phonological store - Which holds the words you hear, like an inner ear 2)An articulatory process - Which is used for words which are heard or seen. These words are silently repeated, like an inner voice, this is a form of maintenance rehearsal.                                                                                                            Visual-spatial sketchpad - Used when you have to plan a spatial task (getting from one room to another, or counting the number of windows in your house). Info is temporarily stored here. Visual info is what things look like and spatial info is the physical reltionahip between things. Logie 1995 divided this group into 1) A visual cache - Which stores info about visual items e.g. form and colour. 2) An inner scribe - Which deals with spatial reltions which stores the arrangement of objects in the visual field.                                                                                                                                       Episodic buffer - Baddeley 2000 added this because he realised the model needed a general store.  The PL and VSS deal with processing and temporary storage of specific kinds of info. The CE has no storage capacity. The episodic buffer also has limited capacity. It sends info the LTM. It intergrates info from the CE, PL and VSS. It also mantains a sense of time sequencing - basically recording events (episodes) which are happening.

Dual task performance - Baddeley 1975 supported the exitstence of a seperate VSS, he showed that participants had more difficulty doing 2 visual tasks ( tracking the light and describing the ltter F) than doing both a visual and verbal task at the same time. This is because both visual tasks compete for the same slave system. This means there must be seperate slave systems.

Evidence from brain damage patients - Shallice and Warringotn studied a man called KF whose short-term forgetting of audiotry info was much greater than that off visual info. His auditory info was limited to verbal materials such as letters and digits but not meaningful sounds (phone ringing). So brain damage restricted to the PL. Trojano and Grossi studied a patient called SC who had generally good learning abilities with the expectation of being unable to learn word pairs that were presented out loud. Damage to PL. Farah studied a patient called LH and found he performed better on spatial taks rather than those involving visual imagery. Suggests seperate visual and spatial systems. However brain damage is traumatic, which may in itself change behaviour so that a person performs worse on certain taks. Individuals may have other difficulties such as paying attention and therefore may undeperform on certain tasks. 

The central executive - Baddeley said the CE is the most important but least understood component of the working memory in 2005, critics feel it needs to be more clearly specified than just being for attention, they beleive that the CE probably has several components, it is more complex than currently represented.

Brain scanning - Braver gave participants a task involving the CE while they were having a brain scan, they found greater activity in the left prefrontal cortex, the activity in this area increased as the tasks became harder since as the demands of the CE increases it has to work harder to fulfil its function. 

Atkinson and Shiffrin 1968.

Sensory register - Where the info is held at each of the senses. The capacity of these registers is very large. The sensory registers are constantly receiving info but most of it receives no attention, a remains in the sensory register for a very breif duration(millisecond).

Attention - If a persons attention is focused on one of the sensory stores, then the data is transferred into STM.

Short term memory - Info is held in STM so it can be used for immediate tasks e.g. working on a maths problem, or remembering directions. STM has a limited duration - info will disappear(decay) quickly if it isnt rehearsed. Info will also disappear from STM if new info enters, pushing out or displacing the original info, since STM has limited capacity.     Maintenance rehearsal - Repetition keeps info in STM but eventually it will create a LTM. Direct relationship - the more it is rehearsed the better it is remembered.

LTM - Has a potentially unlimited capacity and duration. Retrieval is when u get info from LTM which involves the info passing back through STM, then it is available.

Supporting evidence - Controlled lab studies on capacity,duration and coding support the extistence of a seperate short term and long term store. Studies using brain scanning techniques e.g. Beardsley found the prefrontal cortex i active in STM but not LTM, and Squire found hippocampus is active when LTM is engaged.                                   

Case studies - Scoville and Milner studied a man called HM, his brain damage was caused by an operation to remove the hippocampus from both sides of his brain to reduce the severe epilepsy he had suffered.His personality and intellect still remianed intact but he could not form new LTMs but he could remember things from before the surgery.     

The MSM is too simple - It suggest STM and LTM are single stores, however research doesnt support this e.g. WMM which shows STM is actaully divided into a number of different stores in terms of the kind of memory stored. Also for LTM there is different types and each behaves differently e.g. maintenance rehearsal  can explain long term storage in sematic memories but not episodic memories.                                                                                                            

LTM involves more than just maintenance rehearsal - Craik and Lockhart suggested that enduring memories are created by the type of processing you do rather than through maintenance rehearsal, thimgs that are processed more deeply means they are more memorable since more thought is put in to it e.g. you might think what it means and how it links to your knowledge already. Craik and Tulving gave participants a list of nouns e.g. shark and asked questions that involved shallow or deep processing - asked whether the word was printed in capitals (shallow). or or asked whether the word fit into a sentence (deep). The participants remebered the words when they did deep processing - elabrotive processing is key in creating LTMs.

                                                                                                                                  Artificial materials - In everyday life we form memories of peoples faces, names, places, facts and so on. But alot of reserahc studies that provide support for MSM use materials like digits, letters and sometimes words or constant syllables which have no relevance to everyday life so maybe low external validity.

Capacity - The capacity of STM can be tested using a digit span. Jacobs 1887 used this technique to assess STM capacity. He found the average span for digits was 9.3 items and 7.3 for letters. Jacobs suggested this is because there are only 9 digits whereas there are 26 letters.

The magic number 7+-2 - Miller 1956 reviewed psychological research and concluded that the span of immediate memory is about 7 items- sometimes abit more or less. He noted that people can count 7 dots flashed onto a screen but now many more. The same is true if you are asked to recall musical notes, letters and even words. Miller also found people can recall 5 words as well as they can 5 letters - we chunk things together and can remember them more.

*The capacity of STM may be even more limited - Cowan 2001 reviewed other studies and concluded that STM is likely to be limited to about 4 chunks. Research on the capacity of STM for visual info also found 4 items was about the limit (Vogel_. This means the lower end of Millers range is more appropriate.

Individual differences - Jacob found that recall digit span increasd steadily with age, 8 year olds could remember 6.6 whereas 19 year olds 8.6 digits. The age increase may be due to gradual increase in brain capacity and/or it may be that people develop stratergies to improve their digit span as they get older, such as chunking

The size of the chunk matters - Simon found that people had a shorter memory span for larger chunks, such as eight-word phrases, than smaller chunks such as one word syllable words. High validity since results have been confirmed in other research.

Duration of STM - Peterson and Peterson 1959 used 24 students. Each participant was tested over 8 trials. On each trial they were given a constant syllable and a 3 digit number (e.g. THX 512). They were asked to recall the constant syllable after a retention interval of 3,6,9,12,15,18 secs where they had to count backwarsd from their 3 digit number. Particpants on average were 90& correct over 3 secs, 20% over 9 secs and only 2 % after 18 secs. This suggest STM has a very short duration - less than 18 secs as long as verbal rehearsal is prevented.

The duration of LTM - Bahrick 1975 tested 400 people of various ages on their memory of classamtes. A photo recognition test consisted of 50 photos , some from the participants high school yearbook. In a free recall test particpants were asked to list the names they could remeber of those in their garduating class. Particpants who were tested within 15 years of graduation were about 90% accuarte in identifying faces, after 48 years this declined to 70% for photo recogntion. For free recall about 60% accuaret after 15 years and dropping to 30% after 48 years.

Testing STM was artificial - Trying to memorsie constant syllables does not truly reflect most everyday memory activities where what we are trying to remember is actually meaninful. However someitmes we do try to remeber phone numbers or post codes so this study does have some relevance to everyday life.                                          STM results may be due to displacement - They were counting numbers in their STM and this might displace/overwrite the syllables to be remembered. Reitman used auditory tones instead of numbers so that displacement wouldnt occur (sound doesnt interfere with verbal rehearsal) and found that the duration of STM was longer. This suggests that forgetting in peterson study was due to displacment and not decay, if it was displacemtn then the study isnt studying duration but capacity.                                                                                             Bahrick had high external validity , real life meaningful memories were studied. However cofounding variables were not controlled such as the fact that participants may have looked at their yearbook photos and rehearsed their memory over the years.

Coding - Acoustic and semantic coding -  The following words are acoustically similar but semantically different: Cat,cab,can,mad,max,man,map. The following words are opposite sematically similar but acoustically different - great,large,big,huge,borad,long,tall. Baddeley 1966 used words lists like the one above to test the effects of acoustic and semantic similarity on STM and LTM. Participant had difficulty remebering acoustically similar words in STM but not LTM, whereas smeatically similar words posed little problem for STM but led to muddled LTMs. This suggests that STM is encoding acoustically and LTM sematically.

Baddeley may not have tested LTM - LTM was tested by waiting 20 minutes and it is questioable as to whether this is really testing LTM.

STM may not be exclusively acoustic -  Some experiements have shown that visual codes are also used in STM e.g. Brandimote found that participants used visual coding in STM if they wre given visual taks(pictures) and prevented from doing any verbal rehearsal in the retention interval (they had to say lalala) before performing the visual recall task.

LTM may not be exclusively semantic - Frost 1972 showed that Long term recall was related to visual as well as semantic catergories and Nelson and Rothbart found evidence of acoustic coding in LTM.

Artificial stimuli - Baddley used artificial stimuli rather than meaningful material. So you should be catious about generalising findings to different kinds of memory tasks.