The aim of this experiment is to examine the effects
of full and divided attention on false recognition using the DRM paradigm. This
study was conducted to confirm previous findings that divided attention elicits
reduced false remember responses to critical lures. 23 participants took part
in a repeated measures experiment that involved remembering word lists in two
conditions. They completed a recognition test after each condition, based on
the remember/know procedure. Results indicated there was no difference in false
memory production for either attention conditions and more false remember
responses were produced. The findings are inconsistent with previous results
and reasons why will be discussed. 

We Will Write a Custom Essay Specifically
For You For Only $13.90/page!

order now













False memories are based on misleading
information producing memory errors. Memory is quite flexible – when we learn
new information, our knowledge base further develops and our memory allows us
to combine this information with what is previously stored. As a result, memory
is fragile as details about one event has the potential to be contaminated with
details of other events. False memories are studied using the
Deese-Roediger-McDermott (DRM) paradigm (Deese, 1959; Roediger and McDermott,
1995), involving a list of words presented to participants associated with
another unseen word, a critical lure. Evidence shows that people are more
likely to falsely recall seeing the critical lure therefore, a false memory is
created. The Activation Monitoring Theory (AMT; Roediger, Watson, McDermott,
& Gallo, 2001) can explain this. When word lists are encoded, these words
activate others and this causes spreading of activation. Consequently, some
words become incorrectly activated due to association
with others previously stored. At recall, there are memory errors regarding the
source of the memory leading to participants unable to discriminate between presented
words and ones generated internally. Subsequently, this causes participants to
falsely recognise critical lures that were not originally presented to them. The
‘Remember/Know’ procedure (Tulving, 1985) can be used to test recognition of
word lists and involves remember (R) responses or know (K) responses. R
responses are based on conscious recollection whereas, K responses depend on
familiarity feelings. Roediger and McDermott (1995) used this procedure and
found more R responses were associated with false recognition.  

Numerous studies have explained how false memories are
created using the AMT. For instance, researchers looked at factors such as
strength of word lists and presentation speed. Research on false memory and
attention has found evidence of impairing associative activation during
encoding. For example, Knott and Dewhurst (2007) used the DRM paradigm and
manipulated attention. During the full attention (FA) condition, participants studied
word lists silently, while in the divided attention (DA) condition, they had to
generate random numbers simultaneously viewing word lists. Findings revealed
more false R responses were produced during the FA condition, indicating FA
significantly increased the number of false recognition responses to critical
lures. At retrieval, the DA condition had a reduced number of correct and false
R responses, whereas DA had no effect on K responses. Further supporting
research for attention and false memory is put forward by Dewhurst, Barry,
Swannell, Holmes, & Bathurst (2007). They had participants randomly
generate numbers or perform a digit-monitoring task. The results demonstrated
that DA at study produced a reduced amount of correct and false R responses.
Findings implied DA had no significant effect on correct and false K responses.
Research conducted by Dewhurst, Barry, & Holmes (2005) found the same
pattern of results, suggesting DA disrupts participants from activating
semantic associates, thus reducing R responses.

The purpose of this experiment is to replicate previous
research regarding the effects of attention on false recognition. The aim is to
investigate the effects of full and divided attention on false recognition
using the DRM paradigm. Previous research found that DA reduces the ability to
activate critical lures. Based on this, we predict that less false memories
will be generated during DA as words are less likely to be encoded during the
study phase. Additionally, attention will have a more significant effect on R
responses as this relies on conscious recollection while DA will produce less R
responses because there will be a disruption in activating semantic associates
during encoding.




Twenty-three participants were used with an age range
of 15-38 (M = 22.09, SD = 5.60). Inclusion criteria for this
experiment was that participants had to be under 40 years of age. An opportunity
sample was utilised, consisting of 6 male and 17 female undergraduate students
attending City University London as well as family members. Participants were
not paid for taking part in this research.


A 2×2 repeated measures design was used. The
independent variable (IV) was attention containing two levels – full and
divided. The dependent variable (DV) was false recognition rates as well as recollective
experience associated with these false responses, and this was measured by R/K
responses made to critical lures. To reduce the risk of confounding variables
Backward Associative Strength (BAS) for word lists were matched, and
counterbalancing of study lists and attention conditions controlled for order


Participants were given an information sheet and a
consent form before the experiment. A laptop was used to present participants
with the PowerPoints containing the word lists. Two PowerPoints were used each
containing 50 words. The DRM word lists (see Appendix A) were taken from Roediger
et al. (2001). Each word list contained 15 associates of a critical lure. The
first ten words were chosen from ten different word lists which were then
divided into two sets of five, creating two study lists (A and B). The lists
were controlled for BAS using lists with similar means. Presentation of words
in each list was also controlled for, the word with the highest BAS mean presented
first, followed by the other nine words in descending order of BAS means.
Participants were provided with two recognition tests, A and B, which
corresponded to study lists A and B. Both recognition tests contained 20 words
– five critical lures associated with the words lists in each PowerPoint, five
unrelated words taken from any unused word lists, and ten studied words (one
word with high BAS and one with low BAS from each list). The words in the
PowerPoint were presented in the same way for all participants. A booklet of
mazes was also given to participants as a filler task. Furthermore, a 50 beats per
minute metronome audio was used for the DA condition.


Participants reported their age and gender and signed
a consent form before starting the experiment. Verbal instructions were given
before they took part in either the full or divided attention experiment. The
conditions and study lists were counterbalanced. The experiment took place in a
quiet location and lasted approximately 15 minutes. For the FA condition,
participants were instructed to study the words in silence and attempt to remember
as many words as they would be tested on this later. They were then shown a PowerPoint
which contained either study list A or B. The words were visually presented for
two seconds, with a three second break in between each word list. Participants
then took part in a five-minute filler task. After this they were given
recognition test A or B depending on which study list they recently viewed. The
test contained written instructions where participants were instructed to
circle ‘new’ if the words were not familiar. They circled ‘old’ if they
remembered the word. If they circled ‘old’ they either circled ‘remember’ if
they consciously recollected the word or ‘know’ if they felt the word was
familiar (remember/know procedure). Participants were then informed that they
would be taking part in the second condition after a five minute break. For the
DA condition, participants were told that they had to remember as many words as
possible, but while viewing the word lists they must simultaneously generate
numbers from 1-20 at each sound of the metronome. They could not generate
number patterns (e.g. 2, 4, 6…) or count in order of numbers. Due to the
repeated measures design, they were shown the PowerPoint containing the study
list that had not yet been viewed. They then took part in the filler task again
before completing another recognition test relating to the list they studied. After
completion, participants were debriefed and informed of the nature and aim of
the study. An opportunity to ask questions was provided and assured they would be
kept anonymous.



The effects of attention on false recognition
responses to critical lures was analysed using a 2 (attention – full and
divided) x 2 (recollective experience – remember and know) repeated measures
ANOVA (see Appendix B). Attention had two levels (full and divided) and this
was the repeated measures factor. Recollective experience was measured by the R
and K responses and included as a post-hoc factor to examine the effect of
attention and recollective experience. Consequently, the effects of attention
on R and K responses were examined individually. Table 1 shows the Means and
Standard Deviations for the false recognition responses for each level of
attention. There was no main effect of attention, F(1, 22) = 1.78, p >
.05, ?p2 = .08, indicating that there was no significant
difference in false recognition responses to critical lures in either the FA (M
= 0.30) or DA (M = 0.34) condition. However, there was a main effect for
recollective experience, F(1, 22) = 29.85, p < .05, ?p2 = .58. This shows that there were more falsely recognised R responses (M = 0.47) compared to K responses (M = 0.17). There was no significant interaction between attention and recollective experience, F(1, 22) = .98, p > .05,
?p2 = .04.


Table 1. Means
and Standard Deviations of false remember and know recognition responses for
full and divided attention.

























The aim of this experiment was to investigate the
effects of full and divided attention on false recognition using the DRM
paradigm. The results obtained indicate that there was a significant effect of
recollective experience, highlighting that more false R responses were made.
However, there was no significant effect for attention, suggesting there was no
difference in false memory production for full and divided attention.
Furthermore, there was no interaction between attention and recollective
experience, demonstrating that attention type does not affect R and K responses
when making false memories using the DRM paradigm. These findings contradict
previous research as they found a main effect of attention, whereby there were
higher false recognition responses to critical lures in the FA condition (Knott
& Dewhurst, 2007; Dewhurst et al., 2007; Dewhurst et al., 2005). Additionally,
DA had no significant effect on K responses. According to Dewhurst et al.
(2007), this is because K responses are not under the influence of conscious
control, challenging our results which found a main effect of recollective
experience. Additionally, findings revealed that DA produced significantly less
R responses at recognition and this was due to words encoded less accurately during
the study phase. According to the AMT, there was a reduced chance to make
semantic associates with the words leading to the inhibition of R responses. The
present study’s results are inconsistent with this theory and the DRM paradigm.

Previous findings found effects of attention on false
recognition and these were quite robust. The acquired results could be due to
some limitations such as the number of word lists used. For example, Knott
& Dewhurst (2007) used 12 lists per condition (120 words in total), whereas
this study only had 50 words for each condition. This means there was less
opportunity for participants to create false memories as there were less words
to remember. Additionally, the power is likely to be low in our study as we
only tested 23 participants. Knott and Dewhurst (2007) and Dewhurst et al.
(2007) both tested 48 participants, therefore the results obtained may have
been significant if there was sufficient power. Also, the random number
generation (RNG) task for the DA condition may not have been challenging,
therefore future research may want to use varying levels of difficulty, consequently
preventing encoding at the study phase so that semantic associates are not
formed. Further to this, a repeated measures design was used so participants
may have guessed the aim of the study. As a result, they may have not carried
out the RNG task properly, focusing on memorising the words instead. Future
research could perhaps use a between-subjects design to test attention and false
recognition. Furthermore, each recognition test contained only 20 words including
5 critical lures, so participants may have been able to discriminate easily between
old and new words. The odd number of critical lures affected the percentages
when scoring the data. Ideally, each test should contain an even number of
critical lures which was evident in previous research, and at least 36 words, which
was the amount used in Dewhurst et al.’s, 2007 study. A final limitation is
that word lists were used to create false memories, which is artificial stimuli
and not relevant to real-life events such as false memories in eyewitness

In conclusion, this study is inconsistent with previous
findings which found that attention has an effect on false recognition using
the DRM paradigm. This is explained using the Activation Monitoring Theory
which can be used to impair associative activation.