Research Methods Assignment PSYC2013
Investigating Speech vs Non-Speech Sounds on Memory Recall
This research was conducted to investigate further into the effects of sound disruption on
learning and the short memory process and thus aiming to contextualise this to the study of
students. The aim comes from how there are students who have an interest with whether or not
listening to music helps with memory retention while studying. Previous research has been
conducted in this area to that does suggest that irrelevant sounds played while studying hinders the
memory process, a particular study that highlights this is the work of Jones et al (1992). This study in
particular focuses on speech vs non-speech, such as different sine tones unrelated to the study and
sounds affecting the current memory task. In our research the study is measuring the number of
errors made by the participants during a presented memory task under differing conditions. The
main hypothesis going into this study was that the presence of the irrelevant speech and sine tones
during the memorisation phrase will result in a higher number of errors in recall phase when
compared to the control variable. The alternate hypothesis to this is that no clear difference can be
noticed, and any result is purely chance related. However, the main assumption made prior to the
experiment was found to be true with a clear significant value upon analysis so the alternative can
be rejected, this conclusion is made clear with the examination of the ‘Greenhouse-Geisser’ output
and when comparing the conditions side by side in the ‘Pairwise Comparisons’.
In previous years there has been an interest into the effects of irrelevant speech and sine
tones on the ability to recall from the short term memory. Past studies have been conducted
investigating this idea, three key pieces of research that established this idea and served as a basis
for the study conducted here are the ones done by Jones et al (1992) which looked primarily at how
irrelevant speech affected short term memory recall, they were able to conclude there was a
significant affect to declare that irrelevant speech caused disruption. Jones et al (1993) on the other
hand explored the ideas of irrelevant tones affecting the short term memory process, again this
study found there was an affect caused by irrelevant tones. LeCompte et al (1997) conducted a 5
test study looking at both irrelevant speech and irrelevant tones, in particular even testing Jones’
(1992) findings which they concluded to have found that both forms of sound negatively impact the
ability to recall short term memory thus suggesting interruption of the memory process. These
studies all followed the premise of a simple memory task is conducted that required short term
memory use while an irrelevant sound is played, and the performance of participants is recorded. A
final piece of literature to take note of in regard to this field is the journal by Banbury et al (2001) as
this also highlights there is findings to suggest audio can disrupt the workings of short term memory,
which it takes into with further detail. Our study follows similar methodology to these previous
studies workings as a repeat of the aforementioned studies to solidify these findings but also to
contextualise these findings to studying students through the use of a student participant group, our
study also allows for a more generalisable result due to a much larger sample being used.
The conducted experiment used a within participants design so that the same participants
could be exposed to all conditions within the study. The sampling process for the study was a use of
opportunity sampling to recruit participants from the basis that they registered psychology students
that were aged of 18 years old and above, in total 285 people were involved, and these participants
were kept anonymous thought out the study. The participant group was varied with age and sex as
, Research Methods Assignment PSYC2013
there was a mix sex with a majority female sample of 171 to 114 males and with the age range of 19-
44 years old. Before the study took place the participants were given consent forums and general
information about the study ensuring that they were well informed of what was going to happen
and with the knowledge that they could drop out at any point if they choose to do so. For the
procedure of the study participants were asked to complete a task, created with the ‘Pavolvia’
program, to remember the positions of a sequence of letters to which they were then asked the
position of a specific letter. Throughout the experiment different conditions involving sound
happened, the sequence contained the following letters: F, M, T, Q and R. Participants were told to
answer with “Y” (yes) or “N” (no) when asked the position of a letter, the format of which was “R2?”,
this stood for “was the letter R in the second position of the sequence?”. The conditions were as
follows; a control where no sound was played while the sequence was presented, the second was a
condition in which a non-speech related sound was played that varied in sine tones while
memorising and the final was where the sound played was spoken letters, but the letters spoken
didn’t correspond to the ones that were in the sequence. At the end of each sequence in all
conditions before the question was asked a single “beep” sound was played to indicate the end of
the sequence. In total 60 trials were completed with 20 trials for each condition. The experiment did
do a trial run before starting the actual experiment, no data was collected here, and it only served as
a practise for participants to familiarise themselves with the experiment. After completing all the
conditions, the participants were given a debrief form informing further on the study and reinforcing
that the participants are being kept anonymous with raw data being destroyed. The debrief also
contained contact information for support if any participant was harmed psychologically due to the
experiment. Data was then collected and analysed, there was 3 types of response those being
‘correct’ (“Y”), ‘incorrect’ (“N”) and ‘no response’. The ‘incorrect’ and ‘no response’ were both
counted as errors during collection, as the focus was on errors made so these scores were the ones
used in analysis. The data was recorded in ‘Microsoft Excel’ but then transferred over to ‘SPSS’ for
the analysis.
Results suggest that the hypothesis is to be accepted and not rejected for the alternate of no
effect is present. Upon looking at the data analysis from the conditions each condition differed with
the mean and the standard deviation; the ‘control’ had the lowest mean of 5.74 (SD = 2.424),
followed by ‘sine tones’ having mean of 6.28 (SD = 2.960), and finally ‘speech’ with the highest
record mean of 7.81 (SD = 3.043). The total of the mean errors can be seen in fig 1, from this it can
be seen that the control produced the lowest error count. The results from the study do show that
of the hypothesis made can be accepted as true showing a significance within the data. The data was
analysed using a one-way repeated measures ANOVA, the assumption of ‘Sphericity’ was violated so
the significance value was taken from the ‘Greenhouse-Greisser’ which was p<.05 showing that the
results are valid and support the hypothesis. Analysis of ‘Pairwise Comparisons’ also shows that each
condition gathered a p<.05 that there is significant difference between each condition, as seen in fig
2 in more detail.
Investigating Speech vs Non-Speech Sounds on Memory Recall
This research was conducted to investigate further into the effects of sound disruption on
learning and the short memory process and thus aiming to contextualise this to the study of
students. The aim comes from how there are students who have an interest with whether or not
listening to music helps with memory retention while studying. Previous research has been
conducted in this area to that does suggest that irrelevant sounds played while studying hinders the
memory process, a particular study that highlights this is the work of Jones et al (1992). This study in
particular focuses on speech vs non-speech, such as different sine tones unrelated to the study and
sounds affecting the current memory task. In our research the study is measuring the number of
errors made by the participants during a presented memory task under differing conditions. The
main hypothesis going into this study was that the presence of the irrelevant speech and sine tones
during the memorisation phrase will result in a higher number of errors in recall phase when
compared to the control variable. The alternate hypothesis to this is that no clear difference can be
noticed, and any result is purely chance related. However, the main assumption made prior to the
experiment was found to be true with a clear significant value upon analysis so the alternative can
be rejected, this conclusion is made clear with the examination of the ‘Greenhouse-Geisser’ output
and when comparing the conditions side by side in the ‘Pairwise Comparisons’.
In previous years there has been an interest into the effects of irrelevant speech and sine
tones on the ability to recall from the short term memory. Past studies have been conducted
investigating this idea, three key pieces of research that established this idea and served as a basis
for the study conducted here are the ones done by Jones et al (1992) which looked primarily at how
irrelevant speech affected short term memory recall, they were able to conclude there was a
significant affect to declare that irrelevant speech caused disruption. Jones et al (1993) on the other
hand explored the ideas of irrelevant tones affecting the short term memory process, again this
study found there was an affect caused by irrelevant tones. LeCompte et al (1997) conducted a 5
test study looking at both irrelevant speech and irrelevant tones, in particular even testing Jones’
(1992) findings which they concluded to have found that both forms of sound negatively impact the
ability to recall short term memory thus suggesting interruption of the memory process. These
studies all followed the premise of a simple memory task is conducted that required short term
memory use while an irrelevant sound is played, and the performance of participants is recorded. A
final piece of literature to take note of in regard to this field is the journal by Banbury et al (2001) as
this also highlights there is findings to suggest audio can disrupt the workings of short term memory,
which it takes into with further detail. Our study follows similar methodology to these previous
studies workings as a repeat of the aforementioned studies to solidify these findings but also to
contextualise these findings to studying students through the use of a student participant group, our
study also allows for a more generalisable result due to a much larger sample being used.
The conducted experiment used a within participants design so that the same participants
could be exposed to all conditions within the study. The sampling process for the study was a use of
opportunity sampling to recruit participants from the basis that they registered psychology students
that were aged of 18 years old and above, in total 285 people were involved, and these participants
were kept anonymous thought out the study. The participant group was varied with age and sex as
, Research Methods Assignment PSYC2013
there was a mix sex with a majority female sample of 171 to 114 males and with the age range of 19-
44 years old. Before the study took place the participants were given consent forums and general
information about the study ensuring that they were well informed of what was going to happen
and with the knowledge that they could drop out at any point if they choose to do so. For the
procedure of the study participants were asked to complete a task, created with the ‘Pavolvia’
program, to remember the positions of a sequence of letters to which they were then asked the
position of a specific letter. Throughout the experiment different conditions involving sound
happened, the sequence contained the following letters: F, M, T, Q and R. Participants were told to
answer with “Y” (yes) or “N” (no) when asked the position of a letter, the format of which was “R2?”,
this stood for “was the letter R in the second position of the sequence?”. The conditions were as
follows; a control where no sound was played while the sequence was presented, the second was a
condition in which a non-speech related sound was played that varied in sine tones while
memorising and the final was where the sound played was spoken letters, but the letters spoken
didn’t correspond to the ones that were in the sequence. At the end of each sequence in all
conditions before the question was asked a single “beep” sound was played to indicate the end of
the sequence. In total 60 trials were completed with 20 trials for each condition. The experiment did
do a trial run before starting the actual experiment, no data was collected here, and it only served as
a practise for participants to familiarise themselves with the experiment. After completing all the
conditions, the participants were given a debrief form informing further on the study and reinforcing
that the participants are being kept anonymous with raw data being destroyed. The debrief also
contained contact information for support if any participant was harmed psychologically due to the
experiment. Data was then collected and analysed, there was 3 types of response those being
‘correct’ (“Y”), ‘incorrect’ (“N”) and ‘no response’. The ‘incorrect’ and ‘no response’ were both
counted as errors during collection, as the focus was on errors made so these scores were the ones
used in analysis. The data was recorded in ‘Microsoft Excel’ but then transferred over to ‘SPSS’ for
the analysis.
Results suggest that the hypothesis is to be accepted and not rejected for the alternate of no
effect is present. Upon looking at the data analysis from the conditions each condition differed with
the mean and the standard deviation; the ‘control’ had the lowest mean of 5.74 (SD = 2.424),
followed by ‘sine tones’ having mean of 6.28 (SD = 2.960), and finally ‘speech’ with the highest
record mean of 7.81 (SD = 3.043). The total of the mean errors can be seen in fig 1, from this it can
be seen that the control produced the lowest error count. The results from the study do show that
of the hypothesis made can be accepted as true showing a significance within the data. The data was
analysed using a one-way repeated measures ANOVA, the assumption of ‘Sphericity’ was violated so
the significance value was taken from the ‘Greenhouse-Greisser’ which was p<.05 showing that the
results are valid and support the hypothesis. Analysis of ‘Pairwise Comparisons’ also shows that each
condition gathered a p<.05 that there is significant difference between each condition, as seen in fig
2 in more detail.
