Eyewitness Testimony Accuracy

Science Fair Lucy Potts and Christina Liu

Background Information

Lawyers have been using eyewitness testimonies to help prove their court cases since judicial court cases were invented. Eyewitnesses officially state what they have seen, and many times, criminals are convicted on little other than the eyewitness testimonies. But these days, scientists are arguing about whether this is reliable information or not. Our memories are very fragile and not as perfect as some wish to believe.

Human’s memories are not like a video camera. Our minds actually only remember some details, and the rest is filled in with educated guesses. Many scientists and psychologists say that our memories can be altered by someone planting an idea in our head or by discussing the memory with others. This is called “post-event information effect” or “Schemas”. If we hear the suspect has brown hair, we might question our memory of a suspect with black hair, and think that we actually saw a brown haired man. When the witness discuses what he or she saw with others, they speculate some details that they don’t know so that the memory makes logical sense. For example, the exact time of the event might not be known, or the location, so the witness either makes it up or does her best to remember, using other inferences to get the time and location.

Our memories are stored in our head like food is stored in our bodies. The parts we want or need, our brain keeps. The waste or excess that the brain thinks we don’t need, or could be harmful to the body, is not stored. There are three stages of memory; storage, retention and recall. If the memory is not stored in full, then the retention and recall will not be accurate. This is one reason why eyewitness testimonies are inaccurate. Human brains get rid of traumatic memories, and most of the times the witness is experiencing some kind of traumatic event, like a murder. Their memories have gaps and are therefore inaccurate at times.

Faces tend to stand out in human memories. Our minds have the capability of remembering a face for years. How well we can remember the face though depends on how many links we make with the face. On a quick glance, you are less likely to make many links, but over time they develop. For example, you can go to a ten year high school reunion and remember many faces, but a face you saw in the grocery store once, might not be remembered a week later. This can be a good or bad thing with eyewitness testimonies. If the witness knew the person they saw, then they would be able to recognize the face fairly easily, but if they didn’t know them, then they might not remember the face as accurately. Witnesses tend to change their descriptions of an event slightly when they are in court. The longer the gap between when the crime was committed to the time of the trial, can affect the description. Over time your memory degrades, so if the gap between the two events is long, the eye witness's testimony is less reliable.

When a witness is in court and ready to give their testimony, they might not give the description of the suspect from their memory, but instead the description of the suspect from the picture given to them by the police. This has to do with the witness getting nervous from seeing the defendant, but it is also because they don’t want to be wrong. If their description varies from the suspect, then the jury might think that the suspect is not guilty. Due to past experiments and experiences scientists have confirmed that eyewitness testimonies are unreliable. Most eyewitnesses forget 70% of what the really saw and fill in their story with Schemas. Many lawyers try to argue with the eyewitnesses by bringing in a psychologist and telling the judge that the information provided is undependable. However courts are still requiring it in order to reach a decision. It is still a very acceptable form of evidence, even though most scientists say that it should be treated as trace evidence.


The purpose of this experiment is to test the reliability of eyewitness testimonies.


Are eyewitness testimonies accurate?


If a subject gives a testimony right after witnessing an event, a day after, and a week after the event, then the subject’s testimony will change and become more inaccurate as time progresses because memories become weaker as time passes, and the person forgets more details.
Big image



25 subjects

2 movie clips, 1-5 minutes long

Pencil and a sheet of paper to record remembered details, one of each per test per subject

Questionnaire for each subject, for each trial.

Technological device with ability to go on the internet to watch movie clip on

Parts of experiment

Independent Variable: The time frame in which they took their notes

Dependent Variable: The number of observations they make

Constant: The movie clips, the amount of time seeing it (1), the environment, the materials

Control: The first time they record their observations

Experimental Group: the number of questions correct


Gather a group of 25 subjects. This experiment will be testing the accuracy of eyewitness testimonies.

Have the subjects sit down together and watch the same clip of the movie, once.

After the subjects watch the clip, give each person a sheet of paper and a pencil and tell them to write down every detail they remember from the clip including what the people looked like, clothing, and the scenery. Have them write it in a numbered list form. An example of what to write is, “The boy in the clip was wearing a black mask”. Dialog also can be details. Then give them a questionnaire that asks specific questions about the movie clip and have them answer it to the best of their ability.

Then a day later, the same subjects will be asked to write down on a sheet of paper all the details they can remember from the movie clip the day before, without letting them see it again, or what they wrote the day before. They will again fill out the same questionnaire with the questions in a different order.

After a week has passed from the original day of the experiment, have the same subjects write down what they remember again, still not showing them the movie clip or their previous notes again, and again have them fill out the questionnaire with the questions in a different order.

For each subject, count how many details they remembered, excluding the inaccurate ones, for each of the three tests. Compare each of the tests back to the original test, or the one taken immediately after watching the clip, by subtracting that number from the original, which is the control group. This will give the difference and tell how much the witness remembered overtime. Separately count up the wrong details, and do the same process to see if the number of wrong details increased overtime as well.

Repeat steps 2-6 with the same subjects and a different movie clip for a second trial.

Big image


Before taking the quiz after watching the video, most people believed that they would not remember anything, however most were able to answer at least five questions correctly on the first try. As time progressed, people’s scores got worse or stayed the same, whereas some people had test scores that improved over time. It was also noticed that the vast majority of people answered the same questions correctly and missed the same questions. This was because most subjects would keep putting the wrong answer each time the test was given or they gave a completely different answer, one that was not mentioned in the videos. This is a relation to real life eyewitnesses who “change their stories” each time they are asked. When asked about the bigger details mentioned in the video, those were the questions that were answered correctly over the questions asking about smaller details. However, it was interesting to see people think of alternate answers to questions that we had not thought of. This is due to the fact that some people did pay more attention to other small details that we didn’t think to put in the quiz.



In this experiment, there was no trend found in the number of questions each subject answered correctly on Day 0, Day 1, and Day 7.

The statistical analysis, or the Student’s T-Test done, showed no statistical significance. For a T-Test to show statistical significance, the probability has to be 0.05 or less, and the numbers found in this experiment were all around 0.4 to 0.9, which shows there was no significant difference in the number of questions subjects answered correctly over time. Some of the exact statistical values found were 0.61, 0.47, 0.95, and 0.71. Each of these was a different day compared to another day, in the same trail.

There are several possible reasons that this study showed negative results. One possibility is that the subjects remembered the questions on the quiz over time, so their answers stayed the same, resulting in similar scores. Another possibility is that the period of time between each administration of the quiz was too short to show any significant difference in scores. Finally, there were only 10 questions asked on the quiz, which may not have been enough questions to get a range of scores that would show a difference in memory.

Big image

Statistical Analysis and Calculations


The calculations used in this experiment were grading the quiz to give a score out of 10 and averaging the scores.

To calculate the average scores, the scores were added up and divided by 50, since there were 50 data sets. The T-Test was done on Excel. The average of the T-Tests was found by taking the two T-Test results and dividing by two.

Statistical Analysis

A Student’s T-Test was applied to different sets of data looking for statistically significant trends. Day 0 and Day 1 scores were compared for each trial. This was repeated for comparing Day 0 to Day 7 and Day 1 to Day 7. All of these results were not statistically significant, and the experimenters failed to reject the null hypothesis. This analysis was repeated with combining the data from both trials, to increase the power of the analysis. This combined analysis also showed no statistical significance between the scores on different days.


The hypothesis in this experiment was if a subject gives a testimony right after witnessing an event, a day after, and a week after the event, then the subject’s testimony will change and become more inaccurate as time progresses because memories become weaker as time passes, and the person forgets more details. This hypothesis was not proven. The purpose of this experiment was to see how accurate and reliable eyewitness testimonies are. This purpose was achieved. It was found that there was no statistically significant difference in the memory of the subjects as measured by a quiz over time.

Sources of error

In this experiment, there were many sources of error. The number of times a person had seen the movie clip before affected this study because this helped them to remember some of the questions better than others. The noise level in the room at the time the subject watched the video clip also affected this experiment. The subject could have not been paying attention to what they were watching. This is true in real life eyewitness experiences as well, because there could be other things happening that the witness would be watching instead. Some of the questions asked could have been worded in a way one of the subjects didn't understand it, causing them to get it wrong. The age of the person also might have affected their ability to remember the questions as time went on.


If this experiment were done again, videos of crimes in action, such as the beginning of a cop show, would be used to make the situation feel more like an actual eyewitness account. Also, each subject would be made to give a testimony to the crime they saw happen, each time they were asked, along with the questions. This would help see how testimonies change over time even more. If this experiment were to continue, finding out how the subjects react to a real life eyewitness account would be the next step. The experimenters would have to plan a time that something unexpected could happen while their subjects watched, and then ask them afterwards what happened. A larger number of people could be involved in the experiment as well, to get even more accurate results. When the subjects were asked the same questions over and over again, they seemed to memorize them. If this experiment was done again, the experimenter could ask different questions each time, but this might not give accurate results because one quiz might be harder than the other.


Eyewitness testimonies happen every day, with all kinds of cases and situations. Some criminal cases rely on the eyewitness testimony. Knowing the accuracy of their testimonies can help to see if it is really reliable evidence in court. In this experiment, it was found that eyewitness testimonies are relatively reliable and stay the same over time. The results found can be used to say that when repetitive questions are asked, people tend to give the same answers. Knowing how accurate eyewitness testimonies are can help you decide court cases. If all the evidence you have is one eyewitness testimony, it might not be enough evidence to support the whole case.

Video Clips From Experiment


“Emanuel, Steven. Evidence. 4th ed. New York: Aspen Publishers, 2010. Print.”

“Eyewitness Recall - Expert witness is Forensic Psychology - Expert testimony. “Eyewitness

Recall - Expert Witness in Forensic Psychology - Expert Testimony. N.p.,n.d Web. 26 Sept. 2014. <http://www.campsych.com/eyewitness.html>

“Eyewitness Testimony” Simply Psychology. N.p.,n.d. Web 27 Sept 2014.


“How reliable is eyewitness testimony?” http://ww.apa.org. Np., nd. Web. 27 Sept. 2014.


"The Human Memory - What It Is, How It Works and How It Can Go Wrong." The Human

Memory - What It Is, How It Works and How It Can Go Wrong. N.p., n.d. Web. 27 Sept. 2014.