It is still unclear if flashbulb memories represent a particular type of memory or whether they are pretty much similar to most memories for events.
As well as public events, personal events can also have flashbulb memories. Rubin and Kozin (1984) asked for the three clearest memories to be described by their participants. Accidents or injuries to themselves or those close to them constituted a large part of these memories, whereas others related to sports, animals, love affairs or their experiences when starting college. National events only made up three per cent of the sample taken. Surprise was an important factor in this research, as with other research in this area, as well as the number of time people had gone over it in their minds, or rehearsed the memory.
For this present research Brown & Kulik (1977) gives us good guidelines for questions which will help us show the existence of flashbulb memory. With questions relating to the main types of information that were most likely to be remembered the questionnaire that was designed to be used in this study will be able to either support or refute the hypotheses of the research.
The aim of the research is to show the death of Princess Diana as a flashbulb memory by comparing it to the death of the Queen Mother, which is expected to be a long term memory as it was more expected. The two Royal deaths were chosen as they both involved quite prominent members of our society.
Alternative Hypothesis 1
Recall will be greater for the death of Princess Diana than for the Queen Mother as it was unexpected and therefore represents a flashbulb memory, where as the Queen Mothers death was expected so is stored as a long term memory. (Directional)
Null Hypothesis 1
There will be no difference in recall of either of the events, as they are both Royal deaths. Any differences will be through chance factors.
Alternative Hypothesis 2
Feelings will be greater for the death of Princess Diana than for the Queen Mother as it was unexpected and therefore represents a flashbulb memory, where as the Queen Mothers death was expected so is stored as a long term memory. (Directional)
Null Hypothesis 1
There will be no difference in feelings of either of the events, as they are both Royal deaths. Any differences will be through chance factors.
Method
Design
The study was a field experiment with a repeated measures design, where the participants will take part in both sections of the study and their results compared therefore reducing participant variables and allowing them to be their own control group. There were two conditions; experimental where the questionnaire asks about the death of Diana and control condition which was a questionnaire concerning the death of the Queen Mother. The death of the Queen Mother was expected to be stored as a long term memory as it did not meet the criteria for Flashbulb Memory; therefore that is why her death was used as the control condition. The experiment was counter balanced using A-B, B-A with ½ participants getting Princess Diana followed by the Queen Mother and the other ½ vice verse, to no bias in which section of the questionnaire was completed first and also a good cross-section of ages was used to prevent bias. Other controls consisted of equal number of males and females and a quiet environment with as few distractions as possible in order to make the study as free from extraneous variables as possible.
The independent variable was the events of each death that the participants are questioned about and the dependant variable was the recall to those questions. In hypothesis 1 it will be the total recall of both personal and factual events surrounding the deaths, and in hypothesis 2 the total feelings associated with the events. The significance level set is 0.05 level when applied to the Wilcoxon test. This is to try to get a balance between the chances of making a type 1 or type 2 error. As the participant needed to be naïve as to the true nature of the experiment, they were told that the experiment was into memory before completing the questionnaire and were informed as to the exact purpose of the investigation on completion. This isn’t a potentially damaging form of deception so is acceptable for this research and there is no major information being asked for in the questionnaires so the participants do not need to worry about who reads them, though they can leave them anonymous if they wish to. Therefore this deals with any ethical concerns.
Participants
20 participants were used, 10 male and 10 female, and samples were obtained by opportunity sampling. Participants will not have psychological knowledge, so were naïve to the purpose of the study. All of the participants were aged 16 – 75 years old.
Apparatus
- Questionnaires asking questions surrounding the deaths of Princess Diana and the Queen Mother, with questions regarding the place the death happened, when it occurred, who was present, as well as personal questions as to where the participant was and who they were with, and finally questions to establish their feelings at the time of the event. (See appendix for raw data obtained and sample of questionnaire)
- Scoring chart (see appendix)
- Standardised instructions (see appendix)
Procedure
A questionnaire was designed and standardised instructions produced for the investigation, A pilot study using these resources was undertaken in order to assure that questions were understandable and instructions logical. It also assisted with the hypotheses. Any errors in the questionnaires and standardised instructions were made then, before the main study began, to ensure that they were logical to the participants and would not be misunderstood; the discussion section will pick this up again. The study began by obtaining consent of each participant before explaining what their task in the investigation would be. Participants were then given a questionnaire with standardised instructions asking them to fill in what they remembered and not to guess the answer, and not to ask any questions until the questionnaire was complete when the investigator would answer any queries they had. The standardised instructions were printed at the top of each section of the questionnaire, so that no matter whether a participant answered A-B or B-A the instructions were clear to them. The participants were asked to complete the first ten questions as truthfully as possible, with the first five questions about their personal memories of the deaths and the second five recalling the events surrounding the deaths. They were also informed that the feelings section was to show their personal feelings at the time of the event with regards to vividness, emotionality and importance.
On completion of the questionnaire marks were given, one for each correctly answered question and the marks that the participants gave for vividness, emotionality and importance combined to give an overall score for feelings. The participants were thanked and debriefed as to the nature of the investigation, explaining that the research was into flashbulb memory. The results for the total recall for the two events were then compared, as were the results for feelings. (See appendix for Raw Data.)
Results
Table 1
Statistical results of events for hypothesis 1 comparing
the total recall of the deaths of Princess Diana and the Queen Mother
As there are no least occurring ranks (all are negative) then the T value is 0.
From the table for N=20 at the 0.05 significance level for a directional hypothesis, the critical value is 60 for significant difference to be detected. The calculated value must be less the critical value, in this instance it is, therefore it is significant and the alternative hypothesis can be accepted and the null hypothesis rejected.
Table 2 Mean scores of total recall of events for hypothesis 1
Table 3 Statistical results of events for hypothesis 2 comparing
the feelings at the deaths of Princess Diana and the Queen Mother
As there are no least occurring ranks (all are negative) then the T value is 0.
From the table for N=20 at the 0.05 significance level for a directional hypothesis, the critical value is 60 for significant difference to be detected. The calculated value must be less the critical value, in this instance it is, therefore it is significant and the alternative hypothesis can be accepted and the null hypothesis rejected.
Table 4 Mean scores for feelings for hypothesis 2
Graph 1
Graph 2
The Statistical Treatment of Results
The data was collected by counting the total recall of factual and personal events surrounding the deaths of Diana and the Queen Mother to analyse hypothesis 1 and data collected for hypothesis 2 was counting the total feelings of participants for each of the deaths. Each correct factual answer was given 1 mark and 1 each for each personal memory recalled. These were then totalled to give an overall recall score for each participant. The 3 feelings scores were also added together to give a total emotional score for each participant. The resulting data from both of these calculations was interval level (see appendix).
The test used was the Wilcoxon match pairs test. This was used because of the design and the type of data that was collected. The design was repeated measures, therefore the data was related. The data was originally interval, but was converted to ordinal level by ranking it. Because of these two facts Wilcoxon requirements were met and the test could be used.
The Wilcoxon test looks at the difference between related pairs of values. It ranks these according to absolute size, ignoring the direction of the data. Statistic T is calculated by adding the ranks of the difference in the least frequency occurring direction.
- (Research Methods and Statistics in Psychology – Hugh Coolican (Hodder and Staunghton, 1992).
For hypothesis 1 the critical value of T was 0, as there were no least occurring signs, therefore it was a significant result. Because of this the alternative hypothesis was accepted and the null hypothesis rejected. A level of 0.05 was chosen as a balance between a type 1 and a type 2 error. (See Table 1)
For hypothesis 2 the critical value of T was 0, as there were no least occurring signs, therefore it was a significant result. Because of this the alternative hypothesis was accepted and the null hypothesis rejected. A level of 0.05 was chosen as a balance between a type 1 and a type 2 error. (See Table 3.)
The mean scores were calculated for hypothesis 1 (see Table 2) and for hypothesis 2 (see Table 4). The mean was calculated for each to give a visual display of any differences in the events, and there were no extreme scores that would confound the data.
Graphs were drawn using the data from hypothesis 1 and hypothesis 2 (see Graph 1 and Graph 2). The graphs give a visual display of the results for the events.
The Discussion
The aim of this research was to support the theory of flashbulb memory by comparing the deaths of Princess Diana and the Queen Mother. The death of Princess Diana was expected to be a flashbulb memory as it was unexpected; where as the death of the Queen Mother was expected to be a long term memory as it was of no surprise.
To investigate flashbulb memory comparisons were made between total recall of the death of Princess Diana and the death of the Queen Mother. Emotions at the time of each event were also compared.
After analysing the data for hypothesis 1 by using the Wilcoxon T to analyse the data, the results showed a significant result. Wilcoxon T was also applied to the data for hypothesis 2 and this also produced a significant result.
This means that alternative hypothesis 1, “Recall will be greater for the death of Princess Diana than for the Queen Mother as it was unexpected and therefore represents a flashbulb memory, where as the Queen Mothers death was expected so is stored as a long term memory. (Directional)” was verified, and that null hypothesis 1 was refuted. So from total recall results there is evidence of flashbulb memory at the death of Princess Diana.
It also shows the alternative hypothesis 2, “Feelings will be greater for the death of Princess Diana than for the Queen Mother as it was unexpected and therefore represents a flashbulb memory, where as the Queen Mothers death was expected so is stored as a long term memory. (Directional)” was supported and so null hypothesis 2 was also refuted.
There were no anomalous results to affect the results in this research as all participants responded in the expected direction.
Brown & Kulik (1977), said for a memory to become flashbulb the event must be surprising and have real consequences for the person’s life. The present research will look at whether the details of the scene to be imprinted into the person’s memory, as they believed. As this new research looks at an event that is expected to be a flashbulb memory it is similar to the research of Brown & Kulik. However, unlike Brown & Kulik the present research is a one off questionnaire and is not longitudinal.
Whereas as there was rather a lot of media ‘interest’ in the Queen Mothers death, there was a great deal more after the death of Princess Diana, as it was an unexpected event. This also helps people to remember the facts as they have heard and read them numerous times, whereas many people were disinterested by the death of the Queen Mother, as it had been expected and so did not take much attention. This supports Neisser (1982), but is not what the present research will be looking at, although it could have a bearing on the overall results due to reiteration of the events surrounding of Princess Diana’s death, and the media circus that surrounded it.
The research that was undertaken is most similar to that of Brown and Kulik as it looks at the six areas which they found to be most remembered. In future research however this could be changed, with the same participants filling in a questionnaire in a year’s time or even in 6 years time, so that the time elapsed since Princess Diana’s death has doubled.
It would be hard to generalise the results of this single piece of research, as it only used 20 participants. However, 13 other researchers used 20 participants each to independently conduct the same research, all with a significant result, so this accumulated research can be generalised to show evidence of flashbulb memory.
Prior to conducting the main research, a pilot study was used so that any improvements that were needed could be seen. There were no major problems with the pilot study, but the standardised instructions were slightly reworded so that as little misinterpretation as possible would occur. Questions were also reworded slightly and were set out so that the 5 personal questions came first, followed by the factual questions and finally the feelings score chart. This seemed to be logical in sequence, and so was not confusing to the participants.
The wider implication for the type of research used is that the questionnaire could evoke painful memories, such as the death of a friend or family member who died in a car accident. If this was the case with a participant who took part in the main research, it could affect how much of Princess Diana’s death they remember and their emotional scores. Individuals who suffer from Post Traumatic Stress disorder suffer flashbacks. This could be related to flashbulb memory because as well as public events, personal events can also have flashbulb memories. This means that if one of the deaths relates to the death of someone a participant knew they are more likely to remember more details about it, and have stronger emotionality because of it.
In Rubin and Kozin (1984) when they asked for the three clearest memories to be described by their participants, accidents or injuries to themselves or those close to them constituted a large part of these memories. Therefore an event which causes PTS is likely to be a flashbulb memory, where the images reoccur to the individual subsequently due to the trauma it caused them.
The limitations of this research are the number of participants and the age range. Though all in all 280 people in total took part in the research (although the data from each researcher remained independent, each one provided a significant result) it would be enhanced if these people were evenly spread across the ages, and not quite so random. This would show a more even-handed result, though I do not think that it would make a great deal of difference to the overall result. It would also be helpful to have a more varied cross section of backgrounds and ethnicities to see if this made a difference to the result.
Conclusion – The analysis showed significant results, both in recall and feelings, showing flashbulb memory surrounding the death of Princess Diana, but not of the death of the Queen Mother. Therefore the death of the Queen Mother was stored as a long term memory, as it was accepted more easily with her age, but the death of Princess Diana was stored as a flashbulb memory as it was unforeseen.
References
Brown, R & Kulik, J (1977) ‘Flashbulb memories’, Congition, 5, pp.73-99.
Neisser (1982) Memory Observed, San Francisco: W.H. Freeman
McCloskey et al (1988) ‘Is there a special flashbulb memory mechanism?’
Journal of Experimental Psychology: General,
117, pp.171-81
Conway, M.A (1995) Flashbulb Memories, Hove: Erlbaum
Conway, M.A, et al(1994) ‘The formation of flashbulb memories’, Memory and
Cognition, 22, pp.326-43
Rubin and Kozin (1984) Emotional Factors in Forgetting
Psychology in Focus, David Rice
et al (CPL Publishers)
Wilcoxon matched pairs signed ranks test Research Methods and Statistics
in Psychology, Hugh Coolican
(Hodder and Staunghton, 1992)
Appendices
- Pilot Study
- Sample Questionnaire including standardised instructions
- Analysis and Raw Data with calculations
Pilot Study
Raw Data
Mean score of total recall (Hypothesis 1)
Mean score of feelings (Hypothesis 2)
Considerations
- State whether emotions felt now or then
- People may lie (but would it make a difference
- A degree of guessing possible, as with all questionnaires
- Scoring (some lee-way must be allowed)
- Demand characteristics – counterbalance (i.e. ½ Princess Diana then Queen Mother, ½ Queen Mother then Princess Diana) prevents systematic bias
- Cross section of ages
- Consented research
- Gender balance (10 male and 10 female)
- Where you conduct the study ( is it loud or quiet, distractions, etc)
Independent variable = event
Dependant variable = Amount of recall
Total feelings surrounding the event
Hypothesis 1
Participants will remember more about Princess Diana’s death as it she was young and it was unexpected, therefore representing a flashbulb memory. Less will be remembered about the death of the Queen Mother as she was old and people expected her death, and it is therefore stored as a long-term memory.
Null-hypothesis 1
There will be no difference in recall as the events are so similar. Any differences will be due to total chance factors.
Hypothesis 2
There will be more emotionality around the death of Princess Diana as it was unexpected, than with the death of Queen Mother as her death was expected.
Null-hypothesis 2
There will be no difference in emotions surrounding either event as they are both Royal deaths. Any differences will be through chance factors.
Analysis of Data
The test used was the Wilcoxon matched pairs signed ranks test.
The reasons for using this test were because the data was related as it came from a repeated measures design and because the data is at least ordinal level (the actual data is interval level, but was converted to ordinal by ranking it).
Raw Data
As there are no least occurring ranks (all are negative) then the T value is 0.
From the table for N=20 at the 0.05 significance level for a directional hypothesis, the critical value is 60 for significant difference to be detected. The calculated value of T must be equal to or lower than the critical value of T, therefore in this instance this is the case.
As there are no least occurring ranks (all are negative) then the T value is 0.
From the table for N=20 at the 0.05 significance level for a directional hypothesis, the critical value is 60 for significant difference to be detected. The calculated value of T must be equal to or lower than the critical value of T, therefore in this instance this is the case.
