Science Buddies: "Ask an Expert"

I am completing a project on the Accuracy of Eyewitness Testimony. I did surveys to see what "witnesses" remembered about suspects. I was confident that my independent variable was the suspect or visitor I sent into each scenario and that my dependent variable would be the responses of my "witnesses" and my controlled variables would be my questions and the environment. Tonight, I was doing some reading on this site and ran across something that says:

The Independent Variable for Surveys and Tests of Different Groups

When a scientist performs a test or survey on different groups of people or things, those groups define the independent variable. For example:


Question Independent Variable (What I change)
Dependent Variables (What I observe)
Controlled Variables (What I keep the same)
Question: Who listens to music the most: teenagers or their parents? Independent Variable: The groups receiving the survey: teenagers or parents Dependent Variable: The amount of time that each person listens to music per day measured in hours Controlled Variables: Ask the question in exactly the same way to each individual

Help - I am so confused!

Hello. These terms can definitely be confusing. Think of your independent variable being something that CREATE AN EFFECT on your dependent variable. The independent variable will be a category of some sorts. While each person being interviewed has a unique opinion, that's not really a "group" that we talk about when talking about an independent variable. Instead, multiple people being interviewed is known as having multiple samples.

Do you have categories of people that are being interviewed, like boys v. girls, adults v. children, ??? This is something you should consider, as not everyone will perform the same. Boys may have different performance than girls, younger may perform better than older. Think about it....does everyone in your school get the same exact grades or perform the same in sports?

Let's look at an example to hopefully understand the terms better. Say you want to test to see who can put a puzzle together faster, and you want to see if boys or girls put the puzzle together faster. You always want at least three data points per group, so that your experiment is hopefully representative of the true outcome if you were able to test all boys and girls on the planet. If you think about all the boys and girls on the planet, are there other factors that could CREATE AN EFFECT on your outcome? Do you think a 10 year old girl would put the puzzle together just as fast as a 40 year old girl? Do you think times for completing a 100 piece puzzle will be the same for completing a 500 piece puzzle? The answer here is maybe not. So, these are things going on in the background you want to CONTROL, so that your results concentrate on the difference between boys and girls. OK, so from here, we have:

Question: Who puts puzzles together faster? Boys or girls?
Independent variable: gender (boys v. girls)
Dependent variable: time it takes to complete the puzzle
Controlled variables: age, size of puzzle

To control a variable, you have to keep it consistent across all of your samples. So, say you have 5 boys and 5 girls participate. To control the variables of age and puzzle size, you will have to select an age group and puzzle size. For example, test boys and girls between the ages of 10 and 12, and they must put together a 100 piece puzzle. If one boy is 20 while the rest are between the ages of 10 and 12; or if boys complete a 100 piece puzzle while girls complete a 500 piece puzzle, you are no longer controlling these variables.

Does this help? Now try applying this to your experiment, based on what you are changing and testing. Write back with what you come up with, and we can go from there.

Good luck!