Memory

For this section you need to be able to: (a) Describe and evaluate the Levels of Processing framework for memory research (Craik and Lockhart, 1972) and one other theory or model of memory (other than cue dependent theory). The theory we have chosen is the multi store model of memory (Atkinson and Shiffrin 1968) (b) Describe and evaluate in detail one study of memory or forgetting. The study in detail we have chosen is Craik and Tulving’s (1975) Levels of processing study.

The Multi-Store Model

Atkinson and Shiffrin (1968) proposed an early multi-store model of memory which argued that there are three types of information store.

This model suggests that information flows through stores which each have different capacities and durations.
According to this model information is initially stored for a fraction of a second at the sensory organs in sensory memory (also called the sensory register).  Information that is attended to whilst in sensory memory is transferred to short term memory. Short term memory has a limited capacity and duration.  Information that is rehearsed in short term memory is then transferred to long term memory.

Information that is attended to is held in the short term memory for a few seconds and at this stage it tends to be encoded phonetically (acoustically).  The capacity of short term memory is small (approximately 7 items).

Material that is rehearsed remains in the short term memory long enough for it to be transformed to long term memory.  If material is not rehearsed it is lost when new information comes into the short term memory store and displaces it.  Information that is lost through displacement is lost on the first-in-first-out principle.

Long term memory, it is argued, has an unlimited capacity and therefore information can last a lifetime.  Information in long term memory is stored semantically, that is according to its meaning.

Here is an evaluation of the multi store model

Levels of Processing Framework for Memory Research

Craik and Lockhart (1972) developed an influential framework to explain how memory works which is often referred to as the levels of processing approach.  In particular this approach explains why some information is remembered better than other information.

Craik and Lockhart argued that how well information is remembered depends directly on how it is processed.   The levels of processing approach therefore believes that memory is a by-product of the information processing that occurs when attending to information.

Unlike the multi-store model of memory Craik and Lockhart argued that there are two main types of rehearsal needed to transfer information to the long term memory.  The first type of rehearsal is maintenance rehearsal which, like rehearsal used in the explanation of the multi-store model, means going over and over the information.  However, they note that there is a second type of rehearsal called elaborative rehearsal which involves recoding the information in some way.

Craik and Lockhart believe that simply rehearsing information (maintenance rehearsal) is not sufficient to account for long term memory and proposed that it is the level (or depth) of the processing of information that determines the durability of the memory.

They suggested that there are three levels at which information is processed – structural, phonetic and semantic processing.  Structural processing is the shallowest type of information processing, whereas semantic processing is the deepest.

Structural processing involves processing information for what it looks like, for example, whether a word is written in capital letters.

Phonetic processing involves processing information for what it sounds like, for example, does a word rhyme with something else.

Semantic processing involves processing information for what it means, for example does a word have the same meaning as another word.

Semantic processing is the deepest form of information processing because it involves the most cognitive effort and therefore information processed at this level should be well remembered.

Structural processing is the shallowest form of information processing because it involves less cognitive effort and therefore information processed at this level should be less well remembered.

Evaluation of the levels of processing approach

Craik and Tulving (1975) ‘Depth of processing and the retention of words in episodic memory’.

Aim

The aim of this experiment was to test the claims of the levels of processing framework.   In particular the experiment aimed to investigate whether processing words at different levels (structurally, phonetically or semantically) affected the recall of words

Procedure

Craik and Tulving reported several variations of this procedure.

In this procedure the 24 participants were tested individually and were each shown 60 words via a tachistoscope (a devise that allows the researcher to display words for a very short and controlled length of time).  The words were shown one at a time for 200 milliseconds.

Before each word was displayed the participants were asked a question about the word which required a yes/no answer.
Each question related to a particular level of processing.

The first condition involved structural processing, for example, in which participants were asked if a word displayed was written in capital letters.

The second condition involved phonetic processing, for example, in which participants were asked if a word displayed rhymed with another word.

The third condition involved semantic processing, for example, in which participants were asked if a word displayed fitted into a sentence.

The independent variable is therefore the level of processing, that is, structural, phonetic or semantic.

The experiment used a repeated measures design as all participants took part in all three conditions of the experiment.  The experimenters also used counterbalancing as the conditions were run in a variety of orders to ensure that no one type of processing always occurred first or last.

Participants were then given an unexpected recall task in which they had to recognise the test words from a word list of 180 words.  This list was made up of the original 60 words and a further 120 new words.

The dependent variable is therefore the number of words correctly recognised.

Results

Semantic processing did lead to a better memory of words.  That is participant were more likely to recognise words from the list if they had been processed at a semantic level.  Participants were less likely to remember words from the list if they had been structurally processed.

Craik and Tulving also noted that participants did take longer to respond to questions that required a deeper level of processing

Conclusion

Craik and Tulving concluded that the experiment supports the prediction made by the levels of processing approach that the deeper information is processed (with semantic being the deepest, then phonetic and then structural being the shallowest) the more likely it is to be remembered.

Evaluation of the Craik and Tulving study

And here is a quick quiz. memory

And here is another one  TheoriesofMemoryQuiz