Deception; the participants were not aware that they were taking part in a psychological study about memory until after the study had taken place in order to reduce the risk of bias in the experiment. Afterwards they were informed and were given the opportunity to withdraw their responses without prejudice. They were told that no names would be recorded and that confidentiality was assured of their result.
Another risk would be if people then realised the experiment tested memory they could think it was to do with intelligence and worry that they would be judged as stupid’ if they did not remember many words. This could cause psychological distress to that person and might confirm personal fears of insecurity causing further damage. This could be controlled by warning the participant what the study was going to be about before they did it so if they felt embarrassed or worried about their result they could make a more informed decision not to take part.
This table shows how many words were recalled by each participant. It shows that the mean average number of words recalled in Condition A was 15.08; 1.59 less than condition B at 16.67. It also allows us to see that 2/3 of the participants recalled more words with classical music playing than they did in silence. Summary table to show the number of words recalled by each participant in condition A and.
This bar chart shows condition A results in red and condition B in green. From this we can see that both the highest and lowest numbers of words recalled were in condition B. This bar chart clearly demonstrates the difference between condition A and B to vary greatly throughout the experiment. This pie chart shows the difference between the average scores to be quite insignificant as there is only a 5 % difference which means that although most of the numbers of words recalled did increase from condition a to condition b there was not a big difference when taking into account what their scores actually were and working out the average.
I also deduced the range to be 12 for condition A, the lowest/ highest values being 8 and 20, and 17 for condition B, the lowest/ highest values being 7 and 24. The median, however, for both was the same at 16.5. This means there was a wider spread list of values for condition B therefore more variation in how many words people remembered than in condition A. However the descriptive statistics used above only describe what has been found.
In order to suggest the probability of achieving the scores that we did, an inferential sign test was used. A sign test was used as the experimental design was a repeated measures design and the data collected was nominal. When we applied our raw data to the sign test analysis a sign value of 4 was achieved. As this exceeds the critical value of 2, for 12 participants, we can suggest that the raw data achieved is more than 5% due to chance factors and less than 95% due to the manipulation of the independent variable which means our original hypothesis is only partially supported.
From the results obtained in this experiment, we can suggest that our one tailed experimental hypothesis of ‘playing classical music during a memory test will increase the number of words remembered’, is only partially supported, in that only 2/3 (66.67%) of participants recall improved when tested with classical music. This was not found to achieve a significant level of probability < 0.05, which means that recall was not 95% or more due to the manipulation of the independent variable i.e. whether or not classical music was playing in the background. Therefore on this occasion we must accept our null hypothesis of ‘there will be no significant relationship between whether or not music is played in the background and how many words the participants remember’ as there was not a high enough probability to suggest that the variation in number of words recalled was a direct effect of the manipulation of the independent variable.
In our experiment we found that classical music did improve performance of memory as 2/3 of the participants recall improved. This disagrees with the findings of Cohen whose study suggested that participants who were exposed to background noise were cognitively impaired. Our experiment was similar to Cohen’s in that it tested the effect of environmental stimulants on recall ability. Cohen, however, tested the participants under exposure to aircraft noise which is much louder and more distracting than a relaxing piece of music such as Mozart. Our findings, therefore, may be more appropriately compared to the findings of Dr. George Lozanov. Lozanov designed a way to teach foreign languages to children in a fraction of the learning time.
He did this by teaching whilst using certain 60 beats per minute Baroque music and found that his students had a retention rate of 92% and an accuracy of 85- 100% after only thirty days. The findings of which reflect our own in that he used classical music, with the typical 60 beat per minute pattern, to improve performance of memory in the recollection of vocabulary in foreign languages. However, the percentage of improvement in his experiment was 92%, whereas ours was only 66.67% and therefore shows that his experiment was more accurate than our own
However, as our participants were young adults and not children, as in the above, direct and reliable comparisons may not be fully appropriate. This could be due to wider social influences upon adults in relation to children for example a young adult would understand the instructions better and therefore might feel under more pressure to perform than a child. This could have adverse effects and make their performance worse or make them try harder than a child would. Due to this ignorance on the child’s behalf, an experiment involving children is often more natural. Also, differences in procedure and experimental design may have led to the difference in findings.
For example Lozanov choose to teach foreign languages to children and it has been proven that children have a higher capacity than adults for learning language, as they are not as set in speaking their own language as an adult is. The differences found may have related to this factor and so had an effect upon the data that we achieved. If my experiment had been carried out on children instead I think I would have gotten more positive results due to the above reasons.
Other limitations include our choice of method; an experiment is not a natural setting as participants would feel pressurised by the fact that their results were being analysed and might have not been able to concentrate on recalling the words. An improvement could have been a test administered by a normal teacher in a classroom environment, where students would be more used to getting a test and might be more relaxed. This would be unethical, however as it is deceptive. The repeated measures design used meant that the list of words in the second experiment had to be changed because the same participants had seen it in the first experiment and therefore might recall more words after looking at it for a second time.
This would have meant the results were inaccurate therefore another list of words was devised. This could however have caused even more problems because, although care was taken not to use more confusing or longer words from one list to another, the experiment was not standardised and therefore direct comparisons could not be made. A matched pairs design would have allowed comparisons to be made between the lists, but not as accurately between the participants, as every participant will be different.
The opportunity based sample that I used meant that there weren’t an equal number of males and females, therefore it wasn’t representative. A better sampling technique would be to use stratified sampling, in which equal numbers of the same sex can be selected. If the target population was larger, the sample would have been more representative but we needed to use similarly aged participants, as there would have been difference in recall ability between, for example, a four and an eighteen year old. Using people from our own class could have interfered with the result because the participants knew who they were doing the experiment for and could be biased to the experimenter. Also, I think that using younger participants would have given a more natural element to the experiment, as they would not ask too many questions, worry about the results or interfere as much with the experiment by not trying their best.
The participants did not seem to be confused by the standardised instructions but they might not have been clear on why they were doing the experiment, which could have had an effect on their performance. The participants were also talked through the experiment as we carried it out and everything was plainly stated to them. A better way of presentation of the words however, might have been to present them on an over head projector, so that timing could be controlled better; in our experiment, participants were in control of turning over the sheet of paper with the list on and therefore could have turned it over before the test started. The paper was only one sheet thick as well so participants might have been able to see the words even when the sheet was face down.
To expand the experiment further, there are many different variables and different aspects of memory which I could test. The serial position effect, mentioned in my introduction, would be an interesting aspect of memory to test as I noticed in the lists of words there were a significant number of words from the start and the end of the actual list, included in the words the participants remembered. We could test this by setting a quota at, say, the first five and last five words and seeing what percentage of the recalled words were within one of these quotas. We could have used a recording of background noise to repeat Cohen’s experiment or used music with words to test the participants.
A really interesting experiment would be to use children against adults in a memory test with the classical music, as I believe children would give a more positive result in relation to my hypothesis. I don’t believe that there would be a significant difference between males and females, if any, but it would be worth considering. We could have tested different amounts of words see if there is a difference in capacity of recall between a list of one hundred words and thirty words; would the participants be put off by the amount of words and not remember as many for the long list as the short? There are clearly many different experiments relating to memory that could be tested.