Lloyd, M.E., Westerman, D.L., & Miller, J.K.
(2003). The fluency heuristic in
recognition memory: The effect of repetition. Journal of Memory and Language,
48, 603 – 604.
The current
experiments were designed to explore a certain aspect of the discrepancy
attribution hypothesis (Whittlesea & Williams, 2000). The hypothesis states that fluency is
interpreted by an individual to be a sign of previous experience if the fluency
for the stimulus being experienced is greater than the fluency that would
normally be expected of that type of stimulus.
The example from Whittlesea and Williams (1998) used words and
non-words, where the non-words that were orthographically regular had a higher
false alarm rate than real words. So, fitting with their hypothesis, subjects
expected their fluency for real words to be high. The orthographically regular words were still fluent, but because
the words were not familiar, this fluency was attributed to the word’s presentation
at study, hence the higher false alarm rate.
The current
experimenters manipulated expected fluency by manipulating presentation
frequency (1 vs. 5) while at the same time attempting to enhance actual fluency
with a priming procedure. The experimenters predict that even though the
priming procedure should affect actual fluency the same in both presentations,
the expected fluency will be higher in the 5 presentations condition; thus, the
repetition prime will not have its effect in the 5 presentations condition like
it will in the 1 presentation condition.
This
experiment was set up to see whether or not subjects that expected a higher
amount of fluency would be influenced by the presence of higher fluency. 46 words were presented, containing 40
targets. This was a between subjects
design with one group being presented with each of the 40 targets once, and the
second group being presented with each of the 40 targets 5 times each, mixed
randomly, but where the same word didn’t appear back to back. Each test item, target(40) or lure(40), were
preceded by a masked prime, either matched or mismatched. The experimenters predicted that the fluency
gained from the priming would have a greater influence on the recognition in
the 1 presentation group. (1a and 1b were only different in length of delay
between study and test, 20 minutes and 48 hours, respectively.)
Table 1
shows the results for both 1a and 1b. As the table shows, the priming
influenced the group that was presented with the items once but didn’t
influence the 5 presentation group.
Experiments
2-4 were done to try and pick out what was going on in Experiment 1 between two
theories dealing with recollection and expectations of fluency. Recollection doesn’t require familiarity, so
things that increase familiarity would have no effect on tasks that operate on
recollection. This could be one
explanation for the results. On the
other hand, maybe the results from Experiment 1 are due to the difference in
expected fluency.
Experiment
2 was the same as 1 except it was a within subjects design, where half of the
items presented during study were presented once and the other half presented 5
times. Results are in Table 2. When compared to the results in Experiment
1, it seems that subjects adopted an expectation of fluency half way between
the 1 and 5 presentation between subjects groups. This still seemed to leave
both possible explanations.
In
these experiments, the researchers took out one possible explanation:
recollection. If the effect was still
there, then you could safely rule out the recollection explanation.
Subjects
were given a ‘subliminal’ study list.
They were told that the words would be presented to them 5 times.
Another group was told that they were going to be presented with the words
once. A training task was done for
Experiment 3 to set up an expectation of fluency. In Experiment 4, this
training task was left out. Table 3
shows the numbers for both experiments.
Experiment 3 showed that priming in the single presentation group had a
significant effect. Priming didn’t have
a significant effect in the 5 presentation group.
Like I
said, Experiment 4 left out the training task and the priming effect for both
the 1 presentation and the 5 presentations groups were the same. When compared to the results in Experiment
3, there was a significant interaction between the prime and the training task.