Abstract
Loftus and Palmer 1974 suggests memory is unreliable due to reconstructive influenced by schema. This IA aims to replicate this study with minor modifications to the (IV) verbs, to just “Smashed” and “collided” for simplicity. The DV is Speed estimates (mph). 20 participants were selected by opportunity sampling for convenience, randomly allocated into 2 conditions reducing researcher bias, then shown 1 car crash video and asked a fixed question with the (IV) verb manipulated. In condition 1, the verb was: Smashed, the question was— “How fast were the cars going when they smashed each other?”. Condition 2’s verb was “collided”. The average speed estimates were 56mph and 49.5mph respectively. Independent measure was used, reducing demand characteristics. Since the calculated value of the Mann-Whitney U test, 41, at a significant level of 0.05 is higher than critical value 27 for one tailed, the results were not significant at 0.05. Suggesting there was less than 95% certainty that the (DV)speed estimates was impacted by the (IV)verbs and more than 5% was due to other factors. Hence, the hypothesis is rejected and the null hypothesis: there will not be a significantly higher speed estimate for the verb “Smashed” than “collided”, is accepted.
Word count: 199
Introduction
According to the cognitive level of analysis, memory is unreliable due to the “reconstructive nature” of memory being influenced by other factors. “Reconstructive nature” is the brain’s active process to make sense of the current world.
Bartlett’s schema theory argues that this is due to people’s ‘efforts after meaning’ as we process information through filters of past experiences and understandings. Bartlett 1932 aimed to investigate how memory can be distorted by schema. Schemas are cognitive structures defined from understandings formed by culture and social influences, helping us orientate around our environment. His hypothesis was that memory is reconstructive and that people store and retrieve information according to past experiences and understandings formed by cultural schemas. British participants were split into two conditions and were told an unfamiliar native American story: The war of ghosts. Participants in condition 1 were asked to repeat the story after a short time and condition 2, after a long period of time. Both conditions were similar in distorting the story to match their culture and past experiences, to make sense of it. Supporting his theory.
Loftus and Palmer 1974 conducted an experiment to investigate how different critical verbs used in a fixed question will affect the speed estimate in eyewitness testimony. The hypothesis is that people’s memory of a complex event can be distorted. A controlled environment laboratory experiment was used. The independent variable were the verbs used. The dependent variable was the Speed estimates (mph). 45 Psychology university students split into 5 groups, were shown 7 short car accident videos ranging 5-30 seconds long. They were instructed to write a report about what was witnessed and were asked a fixed question about the estimated speed of the vehicle (dependent variable): “About how fast were the cars going when they ***** each other?” In 5 different conditions, the verb (independent variable) was manipulated. In condition 1, the verb used was: Smashed, thus the question was— “About how fast were the cars going when they smashed each other?” In turn, for conditions 2,3,4,5, the verbs collided, bumped, hit, contacted were used respectively. The mean estimate of speed for conditions 1,2,3,4,5 were 40.8, 39.3, 38.1, 34.0, 31.8 mph respectively. Where the estimated speeds decreased directly proportionally to the decreasing severity of the verbs used. Supporting a relationship between language and memory, whereby different verbs activated different schemas in memory such that “smashed” created an image of larger severity than “contacted”.
Aim: To replicate this study with minor modifications to the independent variable, verbs used, to just “Smashed” and “collided” in 2 conditions of 10 participants each and only showing 1 car crash video for simplicity.
Directional hypothesis: There will be a significantly higher speed estimate for the verb “Smashed” than “collided”.
Null hypothesis: There will not be a significantly higher speed estimate for the verb “Smashed” than “collided”.
Methodology
Design
The design used for this experiment was independent measure to reduce demand characteristics. The dependent variable was the participant’s estimate speed (mph). The independent variable were the verbs used, “smashed” and “collided”. At the end of the experiment they were debriefed, allowed access to the results and given the opportunity to withdraw their results. Random allocation was used to avoid researcher’s bias by taking a random number from a box assigning them into either one of the conditions. This allowed convenience. Consent was attained from the participants and they were given the right to withdraw themselves and their results at any point, their confidentiality was also assured to them.
Participants
The target population of this study were students from a private school in Singapore with mostly Singaporeans and a minority of international students. 20 students, 9 males and 11 females aged 16 years and up with the same level of education were recruited. and separated into 2 different conditions. They were generally from middle to upper class families, 2 were Australian, 1 British and 17 southeast Asian, only 14 were native English speakers. Opportunity sampling was used to select the participants as it was convenient as they were in a location at the time of the experiment.
Materials
Approach statement (appendix 1)
Consent form (appendix 2)
Standardized Instructions, given verbally (appendix 3)
Debrief, given verbally (appendix 4)
Questionnaire in a question paper format (appendix 5)
For random sampling:
1. 20 lots (10 for each conditions)
2. 1 Box to store the lots
For the test:
3. 20 question papers
4. 20 pens
5. A stopwatch as a timer
Procedure
Controlled environment: Classroom with desks and chairs are arranged where each participant sits alone at a desk with a pen, flipped over questionnaire and consent form provided. Instruction was stated in the approach statement.
1. Using random allocation, assign 20 participants into two groups of 10 via drawing lots from a box (allocating them into groups—condition 1 or 2).
2. Allow condition 1 into the controlled environment (prepared room) and sit them down at the desks.
3. Read out the approach statement to the participants (asking them to fill in the consent forms) (Appendix 1)
4. Remind them of their rights to withdraw at any time and collect the consent forms.
5. Show participants the video (once only)
6. Allow participants to flip over the answer questionnaire (key word: crashed)
7. Collect questionnaire after 2 minutes.
8. Debrief participants, give them the opportunity to withdraw (Appendix 4)
9. Allow condition 1 to leave the room
10. Reset the room for condition 2
11. Repeat steps 2-9 on condition 2 (with a different questionnaire key word: collided)
Results
Descriptive statistics
Table 1: The mean and standard deviation for conditions A and B
Condition A
(Verb: Smashed)
Condition B
(Verb: collided)
Mean
56
49.5
Standard deviation
29
18
The level of measurement of the data collected is ratio, the mean was chosen as measure of central tendency, as it was the most appropriate. The table indicates the mean and standard deviation for condition A (Verb: Smashed) and condition B (Verb: collided) are displayed respectively. It is seen from the table that mean value in condition A is 56mph, higher than the mean value stated in condition B, 49.5mph. The standard deviation in condition A is 29 compared to condition B that is 18. This indicates a greater variability in Condition A than in condition B.
Bar chart 1 Average speed estimates:
The average speed estimates are 56mph and 49.5mph for “crashed” and “collided” respectively. “Crashed” evidently has a higher average than “collided”.
Inferential statistics
The level of measurement in this study was changed to ordinal data as it did not meet the requirement of a parametric test. The design of the study was independent measure, therefore, Man-Whitney U test was most appropriate to use. The calculated value of the Mann-Whitney U test is 41, at a significant level of 0.05 and number of participants in each condition were N1=10, N2=10, this was compared to the critical value of 27, for a one-tailed. As the calculated value was above this value, the results were found to be not significant at 0.05, which suggests that there was less than 95% certainty that the change in the DV-speed estimates was due to the impact of the IV-verbs and there was more than 5% that this chance was due to other factors. The hypothesis is rejected and the null hypothesis: there will not be a significantly higher speed estimate for the verb “Smashed” than “collided”, is accepted.
Discussion
The results show that the average speed estimates “crashed” was 56mph, higher than “collided” scoring 49.5mph. However, since the calculated value of the Mann-Whitney U test of 41, at a significant level of 0.05 is higher than the critical value 27 for one tailed, the results were not significant at 0.05. Suggesting that there was less than 95% certainty that the DV-speed estimates was due to the impact of the IV-verbs and there was more than 5% that it was due to other factors. Hence, the hypothesis: there will be a significantly higher speed estimate for the verb “Smashed” than “collided”, is rejected and the null hypothesis: there will not be a significantly higher speed estimate for the verb “Smashed” than “collided”, is accepted.
In the original Loftus and Palmer 1974 study, investigating how different critical verbs used in a fixed question will affect the speed estimate in eyewitness testimony, the hypothesis was that people’s memory of a complex event can be distorted. 5 verbs(IV), smashed, collided, bumped, hit, contacted were used. Their mean speed estimates(DV) were 40.8, 39.3, 38.1, 34.0, 31.8 mph respectively. The estimated speeds decreased directly proportionally to the decreasing severity of the verbs used. Supporting a relationship between language and memory, whereby different verbs activated different schemas in memory such that “smashed” created an image of larger severity than “contacted”. However, since modifications were made to the (IV)verbs, to just “Smashed” and “collided” for simplicity, limited analysis could be drawn. Hence, it is not able to support Bartlett’s schema theory stating that we process information through filters of past experiences and understandings due to our efforts after meaning, and that memory is reconstructive and that people store and retrieve information according to past experiences and understandings formed by cultural schemas. It also fails to support his theory that schemas are cognitive structures defined from understandings formed by culture and social influences, helping us orientate around our environment in a cognitively economic way.
The standard deviation in condition A is 29 compared to condition B that is 18. This indicates a greater variability in Condition A than in condition B. This could be due to the fact that some participants are also more familiar with Kmh than Mph due to cultural differences, resulting in some participants giving extreme scores. These outliers evident in the difference in the standard deviation causing variability. If the study were to be replicated, it could be improved by modifying the question, allowing the participants to submit answers in both mph and kmh. Where it will then be converted into mph for calculations and analysis.
Some participants were communicating during the study, in which they could have possibly shared their answers making it a confounding variable. An improvement if replicated would be to move the participant’s seats further away from one another.
Since there were 6 non-native English speakers, language could have been a confounding variable as participants might misinterpret the question due to the lack of understanding. If the study were to be replicated, it could be improved by using another sampling technique such as stratified sampling to ensure that the participants are all native English speakers and are able to understand the question.
The target population of the participants selected from the sample were senior students in a country where its culture is such that majority of them take public transport and are underage to drive Hence, few of them have driving awareness. Thus, it is difficult to generalize the results to the rest of the population. To improve the study, further research could be done in another country with senior students of the same age but who has more driving awareness for comparison.
Some students were also psychology students and hence this might be a confounding variable in the study as they might have guessed the aim of the study and changed their answers. If this were to be replicated, I would suggest another sampling technique could be used to exclude psychology students.
Since this study replicated Loftus and palmer 1974, with modifications to the number of participants from 45 to 20, this small sample size makes it difficult to generalize to the general population. If this study is replicated, a larger sample size should be used.
In conclusion, the stated limitations could have been confounding variables resulting in the value from the Mann-Whitney test to be not significant. Causing the hypothesis to be rejected and accepting the null hypothesis: there will not be a significantly higher speed estimate for the verb “Smashed” than “collided”. If the study were to be replicated, the suggested modifications could possibly improve the study for further research.