Page 1 of 1

Testing the Accuracy of Eyewitness Testimonies

Posted: Sat Sep 09, 2017 10:43 am
by Niyasa
I found this science fair idea on this website and was really interested in it. Although I read through everything on the page, I'm not the best at science and I don't quite understand how to do the surveys or how to even start. Here are my questions.
How many people do I need in a classroom?
How am I going to do this without interrupting a class that needs to work?
Do the participants need to know that I am giving them a survey right after?
Do I tell the participants that some random person is going to walk in so they need to memorize features?
Are survey questions supposed to be open answered or multiple choice?

Thank you for your time. Some of these questions are really stupid, but I'm just really confused...

Re: Testing the Accuracy of Eyewitness Testimonies

Posted: Sat Sep 09, 2017 11:29 am
by cumulonimbus

It sounds like this project is very interesting! To answer your first question, in the "background" section of the project, there is a link about "How Many Survey Participants Do I Need?" : ... ze-surveys

Although there is no single correct number, the amount of people in a classroom depends on the margin of error for which you are aiming. In general, for a smaller amount of error, you want a larger sample size. However, you may be limited by the amount of students that are in the class.

For your second question, you should ask the teacher for permission to conduct the experiment beforehand. That way, the teacher will know not to have any important lesson plans during that time period. If the teacher does not want his or her class interrupted, you can simply find another teacher who is willing.

Thirdly, you should not tell the participants that you are giving them a survey. As the procedure states, you should have the teacher give sealed envelopes to the participants with instructions on when to open them; the participants should not know they are taking a survey until they open the envelopes, as you are testing how well they remember an arbitrary incident, not how well they can recall an incident about which they know they will have to answer questions later.

As for your fourth question, the answer is similar to the previous question: no, you should not inform the participants that a random person is going to enter. As previously stated, you are testing the accuracy of eyewitnesses' memories, people who do not know an event such as a crime is going to happen until they actually see it. In some cases, eyewitnesses may not even realize the importance of what they are seeing, which makes it all the more important that you do not tell the participants to memorize features for a survey later.

Finally, your survey questions should be mostly open-ended, as you want to ascertain what the participants actually remember. If they are given multiple choice questions, they may guess correctly out of pure luck and you will not be certain how much they recall on their own. If you are uncertain as to what types of questions you should ask, the procedure section of the project gives some examples, such as "What time did the person enter the room?" and "What did the person say to the teacher?": ... #procedure
The section also explains some ways to analyze the accuracy of the data you obtain. For further suggestions about formulating survey questions, this Science Buddies link may help: ... n-a-survey

Also, remember that when you are performing projects with human subjects, you will need to obtain permission from the children's parents and teacher. The procedure section of the project has an informative box at the very top explaining the requirements. Additionally, you may want to peruse this link to the requirements for human-related projects, especially if you are considering entering your project in a competition later on: ... egulations

I hope this helps! Have fun completing your project!

Elena Lee

Re: Testing the Accuracy of Eyewitness Testimonies

Posted: Sat Sep 16, 2017 12:24 pm
by Niyasa
Thank you for clarifying and answering my questions! This really helped a lot.
I asked some teachers and they have tight schedules. I decided that instead of having someone come in, I'll just have a teacher show a short 30 second-1 minute movie clip to 2 different groups of 5 students. One group will have a distraction and the other group won't have a distraction.
I got this from the "Make it Your Own" tab.
Would this still be an effective science fair project? It seems like the results are obvious, but I am running out of time.

Re: Testing the Accuracy of Eyewitness Testimonies

Posted: Sun Sep 17, 2017 12:38 pm
by cumulonimbus
That sounds like a legitimate and interesting variation on the project. I would suggest having a larger sample size perhaps, as that would make data trends more clear, but if that is not possible (it sounds like your teachers are too busy to have a larger sample size) then you can work with five. Make sure to provide a thorough analysis of the differences in recall between the two groups before, during, and after the distraction ( i.e. have survey questions about parts in the film clip that occur before, during, and after the distraction). I'm glad I could help and good luck on your project!


Re: Testing the Accuracy of Eyewitness Testimonies

Posted: Wed Sep 27, 2017 3:48 pm
by Niyasa
Thank you so much for your input. It helped me understand it clearer! I will definitely take after your suggestions.

Re: Testing the Accuracy of Eyewitness Testimonies

Posted: Sat Sep 30, 2017 9:52 am
by cumulonimbus
Great! Glad I could help!