Rhodes, M.G., & Kelley, C.M. (2002). The ring of
familiarity: False familiarity due to rhyming primes in item and associate
recognition. Journal of Memory and Language, 48, 581-595.
Under what circumstances will a person attribute the ease
with which they process an item or word to a previous encounter with that
stimuli versus to the context in which the stimuli is presented? This is the
question driving the research reported by Rhodes and Kelley. Jacoby and Dallas
made a simple claim about the attributes that surrounds familiarity. When an
item is processed very easily, meaning that it is easy for a person to read and
or comprehend an item they can attribute this ease of processing to familiarity
with the word. For example, at some level, conscious or unconscious, a person
may be thinking something like this: “I was able to read that word so easily
because I have seen it before or potentially many times before”. Recently, the
concept of attribution of fluency to the past has gained renewed interest. Whittlesea
has suggested that it may not always be the case that fluency perceptual or
semantic may be attributed to the past. In some instances a person will
attribute the fluency with which he or she processed an item to the context in
which it is presented. Although this does not conflict with the original
findings and conclusions of Jacoby and Dallas it provides a fuller description of
the attribution processes that can give rise to feelings of familiarity.
In past research Whittlesea and Williams have manipulated
the ease with which participants were able to process stimuli. They did so by
either presenting a rhyming non-word before the presentation of a word (e.g.,
PINGLE-SINGLE), or a non-rhyming non-word before the presentation of a word
(e.g., BARGEN-PELICAN). Participants were better able to process words (i.e.,
name them) when they were presented after a rhyming non-word than when they
were presented after a non-rhyming non-word. In addition, this effect of
fluency was mirrored in the false recognition. New words that were presented
after a rhyming non-word were slightly more likely to be accepted as old than
were new words presented after a non-rhyming non-word. The idea being tested by
this research is exactly the same as that put forth by Jacoby and Dallas. When
words are preceded by a rhyme it makes it easier to process the word. This ease
of processing is then attributed to the study phase and participants falsely
accept new items proceeded by rhymes as old more often than new item preceded
by non-rhymes.
In a series of four experiments Rhodes and Kelley attempt to
identify the conditions that give rise to feelings of familiarity and the
attribution of fluency to the study environment. For clarification, it should
be noted that Rhodes and Kelley use the term illusion of familiarity to mean
the following: the tendency for participants to attribute an ease of processing
with the previous presentation of an item. Thus, on recognition tests illusions
of familiarity give rise to false alarms. This is an illusion because the word
seem familiar not because it was recently studied (or even something related to
the word) but rather because of the testing context.
Experiment 1
Experiment 1 investigated whether presenting items in
rhyming and nonrhyming pairs at study would increase the chance that fluently
named test items (i.e., rhyming test items) would be judged “old” at test.
Here are the different conditions
Study: Test:
Paired Old
Prime
Rhyme
Rhyme
Non-rhyme
Non-rhyme
Paired New
Prime
Rhyme
Rhyme
Non-rhyme
Non-rhyme
Single Prime
Rhyme
Non-rhyme
At study, participants were presented with 120 study stimuli.
Later they were required to name the test pairs and provide a recognition
judgment for the second word in the pair.
Results
Naming latency:
Old test items were named faster than new test items. Test
items that were preceded by a rhyme were named faster than test items that were
preceded by a non-rhyming word.
Discrimination:
Participants were better able to discriminate between
targets (those words that were presented at study) and distractor items (those
items that were not presented at study) when the test item was preceded by a
non-rhyming prime than by a rhyming prime. In addition, participants that were
presented with single items at study were better able to discriminate between
studied items and distractor items than participants that were presented with
pairs at study.
Response Bias:
Participants were more likely to accept test items presented
with a rhyming prime as old than test items presented with a non-rhyming prime.
When the prime was old and rhymed participants were more likely to accept test
items when than when the prime was new and rhymed.
Discussion
Presenting pairs at study did not decrease the tendency for
individuals to experience illusions of familiarity for test items.
Experiment 2
In this experiment one change was made to the procedure.
There was a condition in which the non-word rhyme was presented but it was not
paired with its rhyming word at study. Instead it was presented with another
word at study and then with its rhyming word at test. Thus, a person could
experience the rhyming non-word at test.
Results
Naming latency:
Rhyming test items were named faster than non rhyming test
items. Old test items were names faster than new test items.
Discrimination and
response bias:
Participants were again better able to distinguish between
targets and distractor items when single words were presented at study than
when pairs of words were presented at study. Participants adopted more liberal
response bias when presented with test pairs that rhymed than when presented
with test pairs that did not rhyme.
Experiment 3
The third experiment added a condition that manipulated the
order of presentation of non-word and the actual word presented at test. The
idea being that presenting the non-word would produce fluency for the actual
word when the pair rhymed.
Results
As can be seen by inspecting Table 5, manipulating the order
of the non-word and word indeed did manipulate the fluency with which
individuals processed the word. When the non-word was presented as a prime the
participants processed the word more quickly. In this experiment participants
were better able to distinguish between targets and distractors when the test
pairs rhymed than when the test pairs did not rhyme. Yet, participants were
more likely to accept items that were presented in rhyming pairs as old than
items presented in non-rhyming pairs.
Experiment 4
For this experiment the researchers presented participants
with rearranged pairs of words and non words. The idea was that participants
might have been processing the pairs as a whole. By presenting rearranged pairs
of words and non-words the influence of the pair as a whole on rates of false
alarms could be investigated.
Results. Discrimination between intact and rearranged pairs
was not as great in the rhyming condition than in the none rhyming condition.
In addition, response bias was greater in the rhyming than in the non rhyming
condition.