Page 172-198, 26 pages Page 1 of7
Chapter 7: Long-Term Memory – Encoding & Retrieval
ENCODING: GETTING INFO INTO LTM
• Encoding: The process of acquiring information and transferring it into LTM, e.g. by rehearsal, by
thinking of other rhyming words, or by using it in a sentence.
• Coding refers to the form in which information is represented, such as a word being coded visually, by
its sound, or by its meaning.
• Retrieval: The process of transferring information from LTM to working memory. One of the main
factors that determines how one can retrieve information from LTM is the way information was encoded
Maintenance Rehearsal & Elaborative Rehearsal
• Rehearsal is the process of repeating information over and over.
• Maintenance rehearsal can keep information in working memory, but is not an effective way of
transferring this information will into LTM = bottom-up
• Elaborative rehearsal is more effective, occurring when you think about the meaning of an item or
make connections between the item and something else you know = top-down
• Craik & Lockhart (1972): Levels of Processing (LOP) theory where memory depends on how
information is encoded, with deeper processing resulting in better encoding and retrieval than shallow
Basics of Levels of Processing
• Depth of processing is determined by the nature of the task during encoding; specifically, memory is
superior when a meaningful connection has been made between an item and something else, e.g. how
useful an object might be on an uninhabited island vs. unmeaningful counting number of vowels
• Shallow processing involves little attention to meaning, with focus on physical features (is word in
lowercase or capitals? Number of vowels?). It occurs in maintenance rehearsal when an item is repeated
without considering its meaning.
• Deep processing involves close attention, focusing on an item’s meaning and relating it to something
else. This occurs during elaborative rehearsal and results in better memory than shallow processing.
• Method: A question is presented, followed by a word, and then the subject’s response. Shallow
processing by asking questions about the word’s physical characteristics, deeper processing about the
word’s sound, and deepest processing involved the word’s meaning. Therefore shallow/deep is more
continuum than categorical
o Shallow: Is the word printed in capital letters?
o Deeper: Does the word [pain] rhyme with train?
o Deepest: Does the word [car] fit into the sentence “He saw a _______ on the street?”
o Subjects are then given a memory test to see how well they recalled the words; deepest
processing is associated with best memory recall.
Difficulty in Defining Depth of Processing
• Describing relative depths of processing can be based on common sense, but sometimes this is quite
difficult, e.g. using a word in a sentence vs. thinking about its utility in a zombie apocalypse, which is
• One way to determine this is to pit the two procedures against each other in a memory experiment by
the method above. Page 172-198, 26 pages Page 2 of7
• Let’s say we find the apocalypse task resulted in better memory; from this we conclude the apocalypze
task results in deeper processing, which allows us to predict memory recall will be better circular
reasoning since depth of processing has not been defined independently of memory performance
• For this reason, levels-of-processing theory not used beyond idea that retrieval is affected by encoding
Research Showing that Encoding Influences Retrieval
• Various types of experiments that vary encoding and measure how retrieval (memory performance) is
Placing Words in a Complex Sentence
• Memory for a word is much better when the word is presented in a complex sentence, e.g. “She
cooked the chicken” not as good as “the great bird swooped down and carried off the struggling chicken”
• The complex sentence creates more connections between the word to be remembered and other
things, which act as cues to help us retrieve the target word.
Forming Visual Images
• Bower & Winzenz (1970): Paired-associate learning in which a list of word pairs is presented for 5
seconds each, then later the first word of each pair is presented with the subject’s task to remember its
• One group was told to silently repeat the pairs during presentation, and another to form a mental
picture in which the two items were interacting; participants in the second group had recall twice as
• Use of visual imagery, images in the head that connect words visually, can create connections to
enhance memory in associative learning
Linking Words to Yourself
• Self-Reference Effect: Memory is better if you are asked to relate a word to yourself.
• Rogers (1977): Subjects presented with a question for 3 seconds followed by a brief pause and then a
word, task to answer “yes” or “no”.
o Physical Characteristic of Word: “Printed in small case?” Word: happy
o Rhyming: “Rhymes with happy?” Word: snappy
o Meaning: “Means the same as happy?” Word: upbeat
o Self-Reference: “Describes you?” Word: happy
• Subjects more likely to remember words that they rated as describing themselves compared to the
other conditions that also elicited “yes” responses – self-reference at 30% compared to 14% for “deep”
• Likely because the words become connected to something the subjects know well – themselves,
therefore resulting in a richer and more detailed representation and thus better memory
• Generation Effect: Generating material actively, rather than passively receiving, enhances learning
• Slameka & Graf (1978): A read group that reads pairs of related words such as king-crown, horse-
saddle. A generate group that fills in the blank with a word related to the first, e.g. king-cr____, horse-
• Subjects then presented with the first word in each pair and told to indicate the second word; subjects
who had generated the second word were able to reproduce 28% more word pairs than the subjects who
just read them.
Organizing Information Page 172-198, 26 pages Page 3 of 7
• Given a list such as “apple, desk, shoe, sofa, plum, chair, cherry, goat, grape, hat”, during recall one
tends to spontaneously group similar items together such as “apple, plum, cherry, grape” and “shoe, hat”.
• Remembering words in a particular category may serve as a retrieval cue to help a person remember
information stored in memory for other words in this category, e.g. apple as retrieval cue for other fruits
• Bower (1969): What if words are presented in an organized way from the beginning? Material to be
learned is presented in “organizational trees” that organize words according to categories, e.g. minerals
grouped into precious stones, rare metals, etc. like chunking seen before
• A group of subjects studied 4 separate trees for minerals, animals, clothing, and transportation for 1
minute each and are then asked to recall as many words as possible from all four. In the recall, they
tended to organize their responses in the same way the trees were, and recalled an average of 73 words
from all four.
• Another group that saw four trees but had it randomized such that each tree contained a random
assortment of minerals, animals, clothing, and transportation. These subjects only recalled 21 words from
all four organizing material into structured groups results in substantially better recall.
• On the other hand, preventing organization would reduce the ability to remember.
• Bransford & Johnson (1972): A nonsensical passage accompanied by an image
describing the passage’s events shown before reading enhanced recall of the
passage, compared to subjects shown the descriptive image after reading or
subjects not shown context.
• The picture provides a context, or mental framework that helps the reader link
one sentence to the next in a meaningful story, so it is easier to comprehend and
• Mnemonic Devices: Anti-chunking, instead of grouping information based on
meaning, we remove meaning to create a retrieval cue. E.g. SOH CAH TOA
• Testing Effect: Being tested on material to be remembered results in better
memory than just rereading it
• Roediger & Karpicke (2006): First, college students read prose passages for
7mins, then solve math problems for 2mins as a break.
• The testing group took a 7mins recall test where they were asked to write down as much of the
passage they could remember in no particular order. The rereading group were given 7mins to reread the
• Then, after a delay of either 5 mins, 2 days, or 1 week, subjects were given the recall test in which they
wrote down what they remembered. With a 5mins delay, little difference between the two groups. With 2
days and 1 week, both groups’ performances decreased; however, the performance of the testing group
dropped much less after the 2-day and 1-week delays
• Words in complex sentence, forming visual image [paired associate learning], self-reference
effect, generation effect, organizing information [retrieval cues], testing effect Page 172-198, 26 pages Page 4 of7
RETRIEVAL: GETTING INFO OUT OF MEMORY
• Retrieval is needed for encoded information to be used, and most failures of memory are failures in
• Retrieval Cues: Words of other stimuli that help us remember information stored in our memory
• Location can be a retrieval cue, as being in a particular place can stimulate memories associated with
that place. Sounds like a particular song, or particular smells can all act as retrieval cues
• In free recall, a participant is simply asked to recall stimuli. In cued recall, the participant is presented
with retrieval cues to aid in the recall of previously experienced stimuli.
• Tulving & Pearlstone (1966): Presented subjects with lists of words to remember, drawn from specific
categories such as birds (pigeon, swallow) or furniture (chair, dresser) although the categories were not
explicitly indicated. Subjects in the cued recall group were asked to recall the words and were provided
with the names of the categories; they performed better at 75% than 40% of free recall group.
• Mantyla (1986): Subjects presented with list of 600 nouns, such as banana and told to write down three
words they associated with each, such as yellow, bunches, tropical. When given a surprise memory test
and presented with the three words they had created, 90% accuracy in recalling the original word.
• If subjects presented with three cues generated by someone else, 55% accuracy. If presented with
cues generated by others without seeing original noun, only 17% -- even if one would expect possible to
infer from clues.
• Retrieval cues are significantly more effective when they are created by the person whose