Zanni & Loftus (1975) used two types of questions. They used the indefinite non leading question, e.g. “did you see a broken headlight?” whereby there is no misleading of information. The other type of question was a definite leading question, e.g. “did you see the broken headlight?” Their study investigated the validity of recalled eyewitness events. This was achieved by asking whether a leading question affects a participant’s response to certain critical words. Participants were shown a reconstruction of a crime and given questionnaires including both leading and non leading questions. This simple change in the independent variable had a noticeable effect in the amount of false responses, the dependent variable.
This study will be an exactly replication of Loftus & Zannis study, however a different video and questionnaires will be used. The aim of my study is to research the effect of leading questions on Eye Witness Testimony. The experimental hypothesis is that leading questions increase the number of incorrectly recalled events. My null hypothesis is that leading questions have no effect on the validity of recalled events.
Method:
Design:
The study was designed as a laboratory experiment with an independent measures design. The sample group was selected into two conditions, a control ‘a’ condition and an experimental ‘the’ condition. The independent variable was the critical word in the question used. This could have been either ‘a’ for the control group, or ‘the’ for the experimental group. The independent variable would have affected the outcome of the dependent variable, in this instance, the number of incorrect facts recalled.
The independent measures design was used to test both groups at the same time, reducing the amount of time needed to conduct the experiment. The laboratory experiment was useful for the study as any extraneous variables could be controlled because it did not affect the results. Noise levels have been found to reduce participant’s concentration in a study. In order to avoid this, the study was conducted during the day in a room in the middle of school. This meant exterior rooms were quiet and noise did not affect concentration levels.
The study involved deceiving the participants about its aim. They believed it was about memory performance and not leading questions. To counter this, a briefing was performed before the start of the study. Participants were informed of their right to withdraw at any time, and they were reassured that any data collected would remain completely confidential.
After the study was conducted the participants were debriefed. The aims of the study were explained. The group were offered the opportunity to withdraw their results. The experimenters contact details were given to the participants. In case they wished to discuss any issues raised by the experiment.
Participants:
The study involved 20 participants, 6 females and 14 males. To ensure a more valid test, none of the participants were psychology students, because they would know the aims of the study and this would compromise the investigation.
This experiment required the use of two conditions with 10 participants in each condition. The group was split into two equal parts. Every alternate person was selected to be in the experimental ‘the’ group, whilst the remainder were selected for the control ‘a’ condition.
The initial sample was chosen by opportunity sampling. The sample was selected from students in a schools common room who were asked if they wanted to take part in a psychological memory test. To ensure the sample was representative, there were three females and seven males in each of the two conditions.
Apparatus/Materials:
In the experiment questionnaires asked leading and non leading questions. There were two separate questionnaires, a control questionnaire that included no leading questions (e.g. did you see a red car?) and an experimental questionnaire that contained five leading questions (e.g. did you see the red car?).
In order for the experiment to be a success, a video was required that the participants could watch and answer questions on. An extract from Crimewatch UK was used. The questionnaire was complied by watching the clip through a number of times. Questions were based on several events that were shown on the video.
To ensure that the method adhered to, a standardised procedure was used. This was a written set of instructions as to how the experiment should be conducted. This ensured the results were accurate.
A briefing was required before the experiment began. This informed the participants that they could withdraw at any time and their answers were strictly confidential. (see Appendix A for the briefing)
After the experiment had been conducted a debriefing was given. The participants were informed that the aim of the study had been to ascertain the effects of leading questions and not to study memory recall. They were also informed that their results could have been withdrawn and not included in the statistical analysis. They were also given a contact number so that they could contact the experimenters if they wished (see Appendix B for the Debriefing).
Standardised Procedure:
The participants were taken to a room that had been prepared with controlled lighting and temperature levels. The room included a television and a video. A briefing was then given, explaining that they could leave at anytime. They were also instructed how to complete the questionnaire.
The Crimewatch UK video was then played which the participants were asked to watch the video carefully. After the video had finished questionnaires were given out, ensuring there was an equal number of control and experimental questionnaires. The participants were then told to complete the questionnaire, in silence and without conferring. The questionnaires were collected and a debriefing given.
Results:
Summary and Descriptive Statistics:
After the data had been collected, a clear trend was apparent. Approximately half of all participants in each condition in the study had one incorrectly recalled fact.
The Graph in Appendix C shows the differences between the two conditions. Condition A ‘a’ and Condition B ‘the’. Condition A was the control condition that contained no leading questions. As the graph shows, the number of incorrectly recalled details was lower. This is because ‘a’ does not suggest an object or occurrence was present in the video. The ‘the’ condition suggested that an occurrence did exist and therefore the number of incorrectly recalled facts was greater.
Inferential Statistics:
The Mann Whitney statistical test was used. This test was employed because the method used was experimental and because the design was an independent measures design in which the two conditions were tested. The data was also nominal and non-parametric. These factors require the use of the Mann Whitney statistical test.
The observed value for the experiment was 45. This was far greater than the critical value of 27 which was at the set level of significance of p<0.05 that was used in this study. This statistical test showed that the results were insignificant.
This meant that the experimental hypothesis can be rejected and the null hypothesis, which stated that leading questions have no effect on the number of incorrectly recalled events, can be accepted
Discussion:
Explanation of findings:
The statistical analysis conducted showed that the effect of leading questions was not as prominent as thought. The statistics show that the results are largely insignificant. Leading questions had little effect on the number of incorrectly recalled events.
There could be several reasons for this. It could be that the impact of leading questions in general has little effect on the amount of incorrectly recalled details. As the statistical conclusion proves, there was a greater standard deviation in Condition A “A” than the Condition B “the” group. There was a greater range of answers in Condition A, but a consistently higher instance of incorrect answers in Condition B.
Relationships to background research:
Two pieces of research were described in the introduction. Firstly Loftus & Palmers (1974) study into the effect of emotionally charged words on the predicted travelling speed of two vehicles involved in a car crash. The method they employed was an experimental method with an independent measures design.
Participants were shown a clip of a car crashes to completed a questionnaire. Some of the participants were in condition A and given emotionally charged words such as collided and smashed. Other participants were in condition B and were given words with less emotionally charged words such as bumped, hit and contacted. This procedure and study design is the identical to this study.
Loftus and Palmer found that the emotional significance as well as its criticality changes the amount of incorrectly recalled facts. For example the word ‘smashed’ was found to develop answers of higher speeds than less emotionally charged words.
The second study that was researched was Loftus & Zannis (1975) study into the effect of leading questions on eyewitness testimony. They also conducted an experiment with an independent measures design. Here, both conditions were tested at the same time. This is exactly the same method employed by this experiment as this study is a replication of Loftus and Zannis study. Their study showed leading questions have a noticeable impact on the on the recall of events.
This study however showed that leading questions did not dramatically affect the number of incorrectly recalled events. A way to improve our results is to repeat the experiment with a larger sample size and under more strictly controlled variables, the results may have supported the findings of Loftus & Zanni.
Limitations and modifications:
There are several variables that could have affected the outcome of the experiment, mainly within the design of the experiment.
Because it was a large sample group (20 people), there was no method available to ensure that the answers the participants gave were completely their own. They could easily look for answers from the person sitting next to them who had a different questionnaire. When the questionnaires had been analysed, it was noted that some participants had put one answer, crossed it out and then written another. Their new answer was the same as the answer from the person sitting next to them. This suggests that some participants may have cheated. This would have distorted the results.
This issue could have been avoided by testing each participant individually. This would have stopped participants looking over their shoulder for the answers. However, this would have been a very laborious process, rendering it unsuitable for the experiment. Another solution would have been to place all the participants in individual cubicles, thus avoiding the opportunity to cheat.
The test was also conducted in a laboratory that lacks ecological (real life) realism. A solution to this would have been to conduct a field experiment instead. This method would have tested the participants in their real life environment instead of a lab situation. However, the variables would not have been as easily manipulated as in a lab experiment.
Another limitation of the study was that the opportunity sampling method employed was highly biased because only participants known to the experimenter were selected. They were not representative of the target population. A solution to this would have been to use random sampling. In this method, more participants than required could have been selected, and then their names drawn out a hat. This would have ensured that the sample was less biased and therefore more representative of the target population.
Another issue was that the video used was not personally relevant to any of the participants. It was a generalised account of events and therefore was not applicable for eyewitness testimonies. Had the video been of a personal nature, the memory trace would have been stronger and therefore less likely to be affected by leading questions.
Implications and suggestions for further research:
This study has some important implications to the real world. Firstly, and most importantly, it suggests that police interviews and questionnaires should be more cautiously constructed. Studies like this one have shown the importance of leading questions, partly helping to develop the Cognitive Interview Technique, used by Police Forces today.
Follow up research would concentrate on whether personal experience to a traumatic event has any effect on the ability to recall a scene of a crime when compared to when it was not relevant.
To conduct this research actors would be used to simulate a bank robbery or another highly stressful situation while people are in the bank. The witnesses would then be asked to describe the robbers using a mix of leading and non leading questions. The answers would then be compared with the answers of the participants who were shown a video of the robbery. They would be asked to describe the robbers using the same questionnaire as used by the witnesses.
This will be useful because it would show the extent to which leading questions alter eyewitness testimonies and whether or not personal relevance is a factor.
Conclusion
The experiment proves that although leading questions can affect memory recall, they do not significantly increase incorrectly recalled details. This is because the video shown was not relevant to the participants and therefore memory traces were not particularly strong. Furthermore no emotive words were used to enhance memory traces. This resulted in the low observed value of 45, which is far from the critical value of 27. This means that the experimental hypothesis can be rejected and the null hypothesis can be accepted, which states leading questions only have a slim effect on society.
References:
Coolican, H.(1994) Research Methods and Statistics in Psychology. (2nd Edition) London: Hodder & Stoughton
Gross, R. McIlveen, R. Coolican, H. Clamp, A & Russell, J (2000) Psychology, a new introduction (2nd Edition) London: Hodder & Stoughton
Loftus, E. F. & Palmer, J. C. (1974). Reconstruction of automobile destruction: an example of the interaction between language and memory. Journal of Verbal Learning and Verbal Behaviour, 13, 585-589.
Loftus, E. F., & Zanni, G. (1975). Eyewitness Testimony: the influence of the wording of a question. Bulletin of the Psychonomic Society, 5, 86-88.
Appendices:
A: Briefing:
Due to the terms of the Sociological Ethical Guidelines, I must inform you all that are you undertaking this study voluntarily and as such you are welcome to leave the experiment at any time.
Also all the data that is collected in this study is completely confidential. You will not have to state your name and we will not share your information with other researchers or institutions.
B: Debriefing
This study has actually conducting how leading questions affect eyewitness testimonies. This is a form of deception, however it was necessary to reduce demand characteristics, whereby you learn the aims of the experiment and change your answers to suit our hypotheses.
During the Study there have been two conditions. Half of you have been placed in the ‘A’ controlled condition, and half in the ‘The’ experimental condition. For example the control group were asked did you see ‘a’ cat, the experimental were asked did you see ‘the’ cat. This was used to test whether a leading question did indeed changed the subjects recall of events.
We are expecting to find that the leading questions with ‘the’ as the key word will produce more yes incorrect answers when compared to the number of yes answers from the ‘a’ questions. This will show that leading questions do affect the recall of events when regarding eyewitness testimony.
Thank you for your time.
C: Graph