Saturday, December 17, 2011

Control and constant variable for a science project?

My problem statement is how accurate are eyewitnesses testimonies?


I basically played out three scenarios and gave them a test on what they remember





my independent variable- The three scenarios


Dependent- How accurate their observations were.





what the he!! are my control and constant?


|||It sounds cool but the problem you are having is because it wasn't really an experiment. There was no hypothesis. It was just an observation.





To make an experiment out of it, you have to have two or more groups observing different things, the characteristics of which, you control. You can be testing anything, but whatever the factor you are testing, that will be your dependent variable. The conditions you set would be the independent variables. I'll give you an example.





Let's say you're testing the accuracy of a person's responses on the basis of stress. You have two scenarios; one which is pleasant and non-stressful and one which is frightening or disturbing. Then you ask two groups of observers the same kinds of questions about each. This can be tricky to set up because of the number of factors you might overlook. But at least now it is an experiment. The control, or independent variable would be stress. Notice, in your present observation, you have no such variable that you are testing.





So you would look at the scores. The more participants, the better. If you tested 20 people and they scored on average, 30% better on the pleasant observation test, then you might have demonstrated a clear relationship.





I mention this example because this very experiment has been clinically conducted many times and it has been shown that there is a clear correlation between the stress level of the observer and the inaccuracy of their answers to questions about it. This has become a precedent in many court cases.





Good luck and have fun. A warning about ethics: do not make your stressful scenario too disturbing. Have it just be an argument or something: no fake gore or anything like that, (even though bloodshed is a SURE FIRE way to blow people's memories all to hell. Don't do it.)





Edit: another thing. Your subjects should not know or even suspect what your experiment is really about. If they know, then the experiment becomes invalid because that would have been a necessary control.





E-mail me if you would like to impress you teacher by double-blinding your experiment. I'll tell you how to do it. That's an important control, too.

No comments:

Post a Comment