Since some people are still doing exams, I’m sharing another academic essay. This one I wrote as a seminar essay with an essay question taken from David Shanks’ PSYC3207 Human Learning & Memory. It’s a great module that I took in my third year of my BSc Psychology at UCL. I believe other degree students can take it as an optional module as well, so if you’re interested in memory and cognitive science at all and have the chance to take it, do consider it. For anyone else – don’t bother reading this tedious academic drivel (that I love so much).
Essay question: Evaluate the levels of processing theory of memory.
According to the levels of processing (LOP) theory of short-term episodic memory, recall of incoming information improves the deeper the information is processed (Craik & Lockhart, 1972). With regards to linguistic information, processing meaning (semantics) is deeper than processing structure (e.g. grammar, graphemics). This essay will examine the evidence for and against the LOP theory and evaluate its account of the memory system. The theory has been convincingly supported by studies demonstrating that semantically processing words leads to longer retention than shallower processing of grammar, critically with rehearsal ruled out as a factor. However, a strong LOP account has been disputed as there is evidence of transfer appropriate processing (TAP). It is concluded that depth of processing does influence recall of information, but that recall is additionally modified by the encoding and testing conditions.
Studies of incidental learning tasks suggest that deeper encoding of word stimuli facilitate subsequent test recall in the absence of rehearsal. Clark and Tulving (1975) presented participants with a series of words and manipulated the type of decision they were to make on each word. In order of increasing depth of processing, participants were to make structural decisions (e.g. whether the word was in capital letters); phonemic decisions (e.g. whether the word rhymed with a specific word); and semantic decisions (e.g. whether the word would fit in a specific sentence). The participants then underwent either a recognition test or a recall test. They were either informed (intentional learning condition) or uninformed about the test (incidental learning condition) from the start of the experiment. Regardless of test type, it was found that memory was better when the meaning of words was encoded than when structure was encoded. This finding was particularly strong, as Craik and Tulving strictly controlled for a number of potentially confounding factors, including “nominal identity of items, preexperimental associations among items, intralist similarity, frequency, recency, instructions to “learn” the materials” and “the amount and duration of interpolated activity”. Hence, only the level of meaning of the stimuli was manipulated. However, there is ambiguity in the definition of “depth”: it was found that the semantic decision task was more time-consuming than the structural decision task. Recall for semantically encoded words was still superior in an experiment where the structural task was more time-consuming (“Is the word brain a CCVVC word?”), however, the conclusions are limited, as time of processing cannot independently measure depth. Nevertheless, Craik and Tulving rigidly established an effect of depth on memory.
Depth of encoding processes as well as rehearsal of information may both influence recall. Crucially, Craik and Tulving found no difference between learning conditions, suggesting rehearsal of the material does not further facilitate recall – otherwise recall would be superior in the intentional learning condition. However, another study disputed the negligent effect of rehearsal: Maki and Schuler (1980) examined the effect of both depth of encoding (i.e. a letter, rhyme or semantic cue) and rehearsal time during a task whereby participants searched word lists for a target word according to the given cue. Although there was an effect of encoding depth, there was an additional effect of rehearsal, whereby longer rehearsal led to better retention. This is contrary to the LOP theory. A limitation of the study was that amount of rehearsal was indirectly inferred from the target word’s position in the word list. Hence, it is not entirely clear whether all participants did rehearse earlier words more than later words. It is nevertheless suggested that depth is not the only facilitating factor of memory.
Examining the spacing effect can help resolve the ambiguity of rehearsal as a facilitator for memory. The spacing effect refers to the finding that spacing out repeated practice sessions of encoded information is more beneficial for memory compared to “massing” the encoding into a single session (e.g. Seabrook, Brown & Solity, 2005). This can be due to spaced information being either rehearsed or processed more deeply than massed information. To resolve this, Challis (1993) conducted a modified experiment similar to Craik and Tulving’s, including both an intentional and an incidental learning condition. In the latter, information was encoded by either semantics (e.g. pleasantness of word) or graphemics (e.g. count ascending and descending letters). After either spaced or massed stimuli presentations, participants were given a graphemically cues recall test. The spacing effect was found in the intentional and incidental-semantic conditions, but not in the incidental-graphemic condition. Similarly to Craik and Tulving’s study, then, rehearsal did not explain the results as incidental learning took place. Indeed, as there was a benefit for semantic encoding, the results supported the LOP account. Challis’ experiment was particularly strong as the design succeeded in distinguishing between the potential effects of rehearsal and depth of processing on an established memory-facilitating effect. This allowed for stronger conclusions to be reached as to the dominating influence of depth.
Eliminating an effect of rehearsal, evidence that depth is not the only factor determining recall has nevertheless been found by support for the TAP account. TAP assumes that recall of information is optimised when the mental processes that occur during encoding match those that occur during retrieval. This is important to consider, as depth at encoding can only be an unambiguous factor provided the type of memory test is irrelevant. Morris, Bransford and Franks (1977) conducted a study similar to that of Craik and Tulving, extended with a critical manipulation: in a second memory test participants were to determine whether a cue word rhymed with a study word. It was found that recall was best for the study words in the rhyme encoding condition than in the deeper, semantic encoding condition. This suggests that it is crucial that the mental operations performed upon stimuli presentation transfer appropriately to the mental operations performed during stimuli retrieval. It should be noted, however, that the overall data indicated an advantage for the semantic study/semantic test condition over the rhyme study/rhyme test condition. The interacting effects of both depth of processing and processing match between study and test therefore remain unclear. Regardless, that shallower processing can sometimes lead to superior recall disputes a strict LOP account.
Considering the difficulty of the recall tests may resolve some of the conflicting implications of the TAP and LOP theories. In Morris et al.’s experiment, there was a significant main effect of retrieval test, whereby the rhyme test was deemed more difficult than the standard yes/no recognition test. This was largely due to the use of novel words in the former and already studied words in the latter. However, there was no significant main effect of encoding task, when – according to the TAP account -, there should be a difference. This highlights a limitation of Morris et al.’s study as they misinterpreted an effect of recall test as an effect of encoding. Marmurek (1995) sought to correct this by an improved replication of the Morris et al. study: at test, participants were presented an unstudied cue word and asked if it related either semantically or by rhyme to a study word that was encoded either by semantics or rhyme. This was a stronger method as all words at test were novel. A main effect of encoding task was established, demonstrating that when test difficulty is controlled for, depth of processing determines recall independently of test type. This supports the LOP theory. It is possible, however, that whether or not there is a clear LOP effect depends directly on the type of test: in a yes/no recognition test, memory was better for words encoded more deeply; whereas in a perceptual identification test presenting both target and new words, memory was better for words encoded more shallowly (Jacoby, 1983). This implies that encoding by shallow reading engages the same mental processes as perceptually identifying words at a later test; and contrastingly, that deeper encoding by generating words engages the same mental processes as recognising words. Thus, rather than merely the depth of a learning procedure providing either good or poor recall, the type of processing at encoding and retrieval – and in particular the match between the two – is what determines whether or not depth affects recall.
In conclusion, the LOP theory of short-term episodic memory is subject to modifying factors such as TAP. Evidence in favour of the LOP theory has been found in experiments which have successfully ruled out rehearsal of stimulus as a confounding factor. There are thus robust demonstrations of better recall the deeper the information is encoded, in particular when test difficulty is controlled for. Importantly, however, depth of processing is not the only factor: the effect of depth is overruled if the mental processes at retrieval appropriately match the mental processes at encoding. A strong LOP account can therefore be discarded. It remains unclear to what extent modifying factors such as TAP, as well as factors not considered here, mediate the effect of depth. There have been some neurophysiological studies finding enhanced activity in the left inferior prefrontal cortex for deep encoding, but not for shallow encoding (Kapur et al., 1994). Perhaps further investigation into the neural correlates of levels of processing and encoding may help resolve the validity of the LOP theory.
Challis, B.H. (1993). Spacing effects on cued-memory tests depend on level of processing. Journal of Experimental Psychology: Learning, Memory and Cognition, 19, 389-396.
Craik, F.I.M., & Lockhart, R.S. (1972). Levels of processing: a framework for memory research. Journal of Verbal Learning and Verbal Behavior, 11, 671-684.
Craik, F.I.M., & Tulving, E. (1975). Depth of processing and the retention of words in episodic memory. Journal of Experimental Psychology: General, 104, 268-294.
Jacoby, L.L. (1983). Remembering the data: analyzing interactive processes in reading. Journal of Verbal Learning and Verbal Behavior.
Kapur, S., Craik, F.I.M., Tulving, E., Wilson, A.A., Houle, S., & Brown, G.M. (1994). Neuroanatomical correlates of encoding in episodic memory: levels of processing effect. Proceedings of the National Academy of Sciences, 91, 2008-2011.
Maki, R.H., & Schuler, J. (1980). Effects f rehearsal duration and level of processing on memory for words. Journal of Verbal Learning and Verbal Behavior, 19, 36-45.
Marmurek, H.H.C. (1995). Encoding, retrieval, main effects and interactions: were
Lockhart and Craik (1990) on the level? Canadian Journal of Experimental Psychology, 49, 174-190.
Morris, C.D., Bransford, J.D., and Franks, J.J. (1977). Levels of processing versus transfer appropriate processing. Journal of Verbal Learning and Verbal Behavior, 16, 519-533.
Seabrook, R., Brown, G.D.A., & Solity, J.E. (2005). Distributed and massed practice: from laboratory to classroom. Applied Cognitive Psychology, 19, 107-122.