kground
Craik and Lockhart (1972) stressed four points in the development of their Levels of
Processing framework. First, they said that memory was the result of a successive series of
analyses, each at a deeper level than the previous one. A shallow level of processing could be
focusing on how a word sounds; a deeper level of processing could be focusing on the meaning
of a word. Second, Craik and Lockhart assumed that the deeper the level of processing, the more
durable the resulting memory. Third, the levels of processing view assumes that rehearsal can be
relatively unimportant. A lot of rehearsal using a shallow level of processing will lead to worse
memory than much less rehearsal using a deep level of processing. The final point had to do with
how memory should be studied: The emphasis is on processing (i.e., what people are doing when
they are presented with the items) rather than on structure (i.e., trying to work out which memory
system is storing the information). Because of this emphasis, Craik and Lockhart suggested that
researchers should use incidental learning rather than intentional learning. The reason is that the
experimenter wants to study the effects of a particular type of processing. If the participant
knows that there will be a memory test and tries to learn the material (i.e., intentional learning),
the participant might use a different type of processing. At the very least, the experimenter will
be unsure of the type of processing used. On the other hand, if the participant is unaware that
there will be a memory test and is not intentionally trying to learn the material (i.e., incidental
learning), there is no reason for the participant to use a different type of processing.
The typical levels of processing experiment used incidental learning. A participant is
asked to rate words based on the number of letters or consonants, or on the words’ pleasantness.
Because the participant is unaware that there will be a memory test later, the experimenter can
assume that once the rating task is over, the participant will not process the item further. After all
the ratings are done, the participant receives a surprise recall or recognition test. The usual
finding is that the deeper the level of processing, the better the performance on the test.
In this experiment, it is expected participants should correctly recognize more words with
the deep processing task and fewest words with the shallow processing task. The reason is that
on a standard recognition task, most people use semantic information as a cue to retrieval, and
the type of processing most appropriate for a semantically-driven task is also semantic
processing at test. The effect is quite robust and is not limited to recognition tests. However, the
effect can be altered if the test is changed. For example, if the test asked you whether there was a
word on the list that rhymed with a particular test item, you would do better for those items you
made a rhyme judgment on in Phase I than the items that you processed deeply.
Experiment details
Participants were all Skidmore college students taking a Cognition class and completed the experiment
for credit. Data were collected from each participant one a computer. All participants self-tested (i.e.,
completed the experiments on their own computers following the instructions.)
Participants received the following instructions:
There are two parts to this experiment.
In Phase I, you will see a word and a judgment task. There are three types of judgment tasks.
PS330: Research Methods in Memory
Mini-Lab Background sheet
1. The first type of judgment is to decide if the word is in UPPERCASE letters. For
example, you might see “Is PEN in uppercase?” You should respond Yes. You might also
see “Is pen in uppercase?” You should respond No.
2. The second type of judgment is to decide if two words rhyme. For example, you might
see “Does PEN rhyme with bus?” You should respond No because they do not rhyme.
You might also see “Does PEN rhyme with hen?” You should respond Yes.
3. The third type of judgment is to answer a question about a word. For example, you might
see “Is PEN a type of vehicle?” You should respond No. You might also see “Is PEN a
writing implement?” You should respond Yes.
There are 60 judgment trials.
In Phase II, you will be shown a series of words, half of which were shown in Phase I, and half
of which are new words. For each word, please answer the question, “Was this word in Phase I?”
There are 120 words shown in Phase II.
Materials
All words used for phases I and II were concrete nouns, three to six letters, and one to two syllables (e.g.,
dog, witch, pond, miner, juice, glove, copper, daisy, lamp, chapel).
Data record
At the end of the experiment each participant received three encoding scores, which represented their
accuracy for each of the three tasks during phase I, and three test scores, which represented their hit rates
for words belonging to each of the encoding task categories.
Research Question(s):
What was the average task accuracy during encoding for each task? Was encoding accuracy different for
the three types of tasks?
What was the average hit rate for each encoding task category during testing? Was accuracy on the test
different for the three types of encoding tasks? Was there a levels of processing effect in which hit rate
accuracy increased as encoding became deeper?