To investigate the effect of misleading questions on the accuracy of witness statements

Eye-Witness Testimony

Aim: To investigate the effect of misleading questions on the accuracy of witness statements.

Introduction: Reconstructive memory, or the idea of a schema, a mental framework into which related information or ideas fit together, was first introduced by Bartlett in 1932. In storing information we may reconstruct it to fit into our own personal memory, so it conforms to our existing beliefs, values and expectations. What we remember is influenced by what we know. Confabulation is the process, when we unconsciously adapt our memories to fit in with our existing experiences.

Our experiment was based on the work conducted by Loftus and Zanni on eye-witness testimony. For example, they did an experiment where participants were shown a film of mulpitle car accidents, and then were asked questions, concerning the accidents. There were 2 groups and they were asked 22 item questions and 6 critical questions. Group A was asked indefinite article questions, for example “Did you see a broken headlight?” and group B “Did you see the broken headlight?”. Indefinite article did not imply as to whether article was present or not and definite article did imply that it was present. The results, 7% of group A said “yes” and 15% of group B said “Yes” to an item which was not there, suggests that the wording of the question can have influence on the answer of the eye-witnesses. So the chance of an incorrect answer increases if a misleading question is asked.

Another experiment has also been conducted by Loftus and Palmer in 1974, where participants were shown a video of a car accident and were asked to judge the speed of the cars. The wording of the question again had an effect on the answer of the participants. The questions were different in the type of verb that was used, i.e. “About how fast were the cars going when they hit?”. For 5 different groups the word “hit” was replaced by smashed, collided, bumped and contacted. The speed estimated was lowest for the “contacted” group – 31.8 mph and largest for ...

This is a preview of the whole essay

Experimental Hypothesis: Asking misleading questions increases the chance of participants reporting seeint items / actions which did not occur (Q2-4) or causes participants to report an increase in the number of people (Q1) or the speed of the car (Q5) as suggested in the misleading question.

Null Hypothesis: There is no difference between answers to straightforward questions and answers to misleading questions.

Method

Design: The independent variable is the straightforward or the misleading question.

The dependent variable is the answer given.

The study was an experiment with an independent groups design (different participants in each condition). The main variable that has not been controlled are the individual differences between people as previous experience and memories could also affect the answers given. Repeated measures to get more reliable results would have been inappropriate because people would be very likely to remember their previous answers, could see the difference between questions and guess what the study is about, due to the demand characteristics of people. That would influence the answers given and so would be unreliable. The ethical issue raised is that fully inform consent was not given to the students before the test, but given after.

Participants: 23 Mixed AS students.

Apparatus and Materials:

- Open University Video “Eyewitness Testimony” from the start for the first 2 minutes

- Straightforward Questionnaire (see Appendix)

- Misleading Questionnaire (see Appendix)

Procedure: The experimenter told the students that they are going to be eyewitnesses to a crime scene. She then showed the student the video of the crime, where two men robbed a bank. The students were then given either a straightforward or a misleading questionnaire. The questionnaires were completed and then collected by the teacher in silence, so that the students would not communicate between each other.

Results:

See the Results table and the Bar Charts.

Discussion: The results did not show any significant trends whatsoever. It is clear from the results that the students were not influenced at all by the way the question was asked. The results do not support the experimental hypothesis; neither do they support any previous research.

However, the results could be unreliable in many ways. First of all, everyone is different. So every tested person would have had different past experiences and memories which could affect their answers. Another important factor is that because the students were not told what the experiment was for, they were very likely to try and guess what the experiment was about, due to demand characteristics, and if they succeeded at guessing that could change their answers. It is not like they are giving evidence against someone, which could have great consequences, but they were in a classroom and whatever the students would answer, it would not have any impact on anything important. There is also an issue that there is no emotional factor involved – seeing a video of a car crash is not remotely similar to seeing it in real life. Emotions can create flashbulb memories, which means the witness could recall the event with very fine details. This issue was also not considered in the two experiments by Loftus discussed earlier.

To improve the experiment, more participants should be tested. Due to the limited amount of students in a classroom, it was harder to do. However, the results would have been much more reliable if more people were tested. To improve the ecological validity of the experiment, the emotional effect should be considered. For example, the experiment could be done on an important public tragedy, such as the Twin Towers, which would involve the emotional effect.

References: Rice,D and Haralambos,M (2000) Psychology in Focus AS Level Causeway Press p.34-46.