Title: Memory in the Real World 1
1Levels of Processing Overview
2Background
Problems with the modal model also lead to the
development of the levels of Processing approach
Craik and Lockhart (1972)
LOP retained distinction between STS and LTS, but
challenged key modal model assumptions
1. Memory involves items being held in structural
stores 2. Fate of items determined by property of
the store 3. Research should determine number and
properties of stores.
The memory trace consists of a record of those
processes carried out for the purposes of
perception and comprehension. The deeper the
processing, the better the retention.
3The LOP Framework
Stimulus processing (of words at least) can occur
at a number of different levels
Type of processing Level of Processing Retention
Orthographic Shallow Poor Phonological
Semantic Deep Good
Existing evidence from incidental learning
paradigms supported this contention
As the subjects are unaware that their memory
will be tested, they presumably encode the words
only as instructed by the experimenter.
4The LOP Framework
According to LOP framework A memory trace is a
record of the analyses carried out during the
conscious processing of new information Therefor
e Memory function should be explored by
investigating how variations in the way
information is processed affect our subsequent
ability to remember it. CL proposed that
processing begins at a superficial and shallow
level, and proceeds to deeper and richer
levels. Shallow encoding leads to poorly
retrievable traces Deep encoding leads to more
durable and retrievable traces
5Evidence for LOP - encoding
Craik, 1977 - Is word in upper or lower
case? - Does the word rhyme with mat? - Does
the word fit the end of a sentence?
6Evidence for LOP - rehearsal
LOP also made a distinction between two types of
rehearsal Type I - Maintenance rehearsal Type II
- Elaborative rehearsal
Only type II rehearsal is associated with
increased retention
Craik and Watkins, (1973) - Subjects listen to
list of 21 words and told to recall last word
beginning with e.g. g
As Ss do not know which g word will be last,
they will rehearse the first one, until the next
one appears.
Thus by varying the number of words between g
words, the amount of rehearsal can be
systematically varied.
Amount of rehearsal had no effect on subsequent
recall
Professor Sandford - read prayers before meals
over 5000 times, but recall was very poor when
tested formally.
7Evidence for LOP - real world
Svenson (1977) - classified students according to
how they read a piece of text.
Surface approach - I tried to memorize
everything, I focused on specific details as
opposed to the whole
Deep approach - I tried to identify the
principal ideas, I focused on the overall
meaning / the authors points
Students who had adopted the deep approach
recalled substantially more in an unexpected
recall test
23 (of 30) students admitted adopting the same
approach to required course reading
9 (of 10) who adopted a deep approach passed the
course
3 (of 13) who adopted the shallow approach passed
the course
8Subsequent Developments
congruency effect Subjects produce better
performance when the initial orienting task
evokes a yes response. (AKA compatability effect)
Figure 2.3 p30 Parkin
This effect is most pronounced for the semantic
orienting task
Yes responses allow potential associations to be
made between the stimulus and its reference frame
9Subsequent Developments
Elaboration Effect Different semantic orienting
tasks lead to different levels of retention
(Craik and Tulving, 1975).
Could the word watch fit the following sentence?
1. She dropped her ______
2. The old man hobbled across the room and
dropped his _____ in the jug
Condition 2 leads to better retention
Sentence verification also leads to better
retention than judging semantic category
Craik and Tulving proposed that the elaborateness
of the processing (the number of associations
with stimulus generated) was important. A
similar point had been made by William James.
10Problems with LOP
Late 1970s - Criticisms of LOP begin (Baddeley,
1978, Eysenck, 1978, Nelson, 1977)
Circularity -
Retention test performance (Semantic better than
non semantic)
Depth of processing Semantic Deep,
non-semantic Shallow
In order to break circularity an independent
measure of depth is required
11Circularity
In view of the vagueness with which depth is
defined, there is the danger of using retention
test performance to provide information about the
depth of processing, and then using the putative
depth of processing to explain the retention
test performance, a self defeating exercise in
circularity Eysenck, 1978.
Loftus, 1980 - How deep is the meaning of life?
Subjects view pictures of scenery A) Judge of
picture taken by sky / count trees (SHALLOW) B)
Rate pictures for pleasantness (DEEP) C) Think
about the meaning of life. (UNKNOWN)
Recognition memory in A worse than in both B and
C, which did not differ Therefore (following LOP
logic) allowing the conclusion that the meaning
of life is relatively deep...
12Other criticisms of LOP
Weak theoretical power
The development of the concepts of elaboration
and compatability weakened the simplicity of the
LOP approach
Both were as unmeasurable as the concept of
depth
Overspill coding Original LOP model required
co-ordinality assumption the nature of
processing undertaken is a reflection of the task
demands.
NOT TRUE Shallow orienting tasks almost
certainly involve some semantic processing
XXXX X GREEN YELLOW
The Stroop effect is an example of unwanted
semantic processing.
13Other problematic data
When is an orienting task semantic?
Hyde and Jenkins (1973) - Considered sentence
frame judging a non-semantic task (and found poor
recall)
Also considered estimating a words frequency to
be a deep orienting task, but not clear why this
should be.
Some shallow orienting tasks lead to excellent
retention
Kolers (1979) Trained subjects to read
mirror-inverted script.
Benefits of training evident 1 year later
14Other problematic data
Amnesic patients
Amnesic patients have no problems performing
semantic orientation tasks - but they still have
poor memory
Maintenance rehearsal can lead to better retention
Nelson (1977) Processing twice at the phonemic
level is better than once - as long as the
rehearsals are spaced.
Maintenance rehearsal facilitates priming of
existing representations, but does not encourage
inter-item associations (which lead to better
long term retention)
15Other problematic data
Several researchers suggested that
distinctiveness of processing was as important
as depth.
Eysenck Eysenck (1980) - subjects read
irregular words regularly - Glove - pronounced
as in Cove
- A shallow orienting task leading to a
distinctive memory
16Other problematic data
Morris, Brandsford Franks (1977) Ss perform
deep (semantic) or shallow (rhyme) orienting
tasks.
Standard recognition or Rhyming recognition test
LOP effect for standard test. But opposite for
rhyming test Deep processing does not always
enhance memory
17Transfer Appropriate Processing
In order to account for their data, Morris,
Bransford Franks developed the concept of
TAP Memory performance depends on the extent to
which processes used at the time of learning are
the same as those used when memory is tested.
LOP approach assumed that semantic processing was
always superior to non-semantic processing
The transfer appropriate processing approach
demonstrates that a form of encoding which is
shallow for one purpose might be deep for
another.
Some researchers argue that TAP can explain
differences between implicit and explicit tests
of memory (see lecture 5)
18The legacy of LOP
- Craik and Lockhart assessed the impact of the LOP
approach in 1990 - They admitted that they were wrong on several
accounts - 1. Shallow processing does not always lead to
rapid forgetting - 2. Processing does not necessarily proceed from
shallow to deep. - BUT pointed out that
- LOP was presented as a framework, not a theory.
- 2. Many criticisms were actually addressed in
later versions of the model.
19The legacy of LOP
1. LOP is a useful rule of thumb - deeper, more
elaborate encoding usually leads to better
retention 2. For this reason, LOP manipulations
have been useful in huge range of studies
(psychopharm, attention, educational psych
etc) 3. Memory researchers now attempt to
understand memory in terms of its processes
rather than its putative static structures 4.
Highlighted the link between perceptual processes
and memory processes 5. LOP Highlighted the
variety of encoding which can occur 6. LOP
approach led to the development of TAP
20Summary