linda
Posts: 1
Joined: Mon Nov 07, 2005 8:13 am

### Control and variables

I am doing a science fair project on the affects of ethylene on fruit ripening....having trouble charting my results b/c I'm confused about the concept of control and variable. I can't figure out which is which

H E L P
I am confused about variables and control. I have to make a bar graph to chart my results.

Former Expert
Posts: 260
Joined: Thu Aug 25, 2005 7:44 am
Occupation: Software Engineer/QA Lead - Quality, Risk Assessment, Statistics, Problem Solving
This is a pretty big topic, but I'll take a stab at it.

To understand these concepts you have to go back to your hypothesis (question).

A hypothesis is framed as:

"If I set up the situation in a particular way, then I predict that a specific thing will occur."

Variables are anything in your "set up the situation" that can change the outcome.

For example, if my hypothesis is "If I feed mice a high carbohydrate diet, they will gain weight". then the variables might be things like:

Age of the mouse
Environment (temperature, light cycles, etc)
Exercise opportunities
amount of the food
macronutrient mix in the food (ie, protien vs carbohydrate vs fat)

Generally the idea in a science fair experiement is to hold as many variables as possible constant, and to very only one thing, to see if your outcome changes. In the above example, you'd want to keep everything the same for all mice except the macronutrient mix in the food.

The concept of a control is a little different.

It is possible when you show a relationship between events that you don't have cause-and-effect. In my above example, the high carbohydrate diet might in fact cause the mice to gain weight, but we have not proven that is the cause.

To prove that some other factor was not the cause, you would want to have some other mice, with all the same variables kept constant except that they got a "normal" balanced diet.

If your "normal" mice also gained weight, in a similar fashion to the "high carb" mice, then your hypothesis is not proved, instead some other factor you did not account for is at work (portions too large?)

So a "control" is when you keep all variables constant, and more importantly "normal". Then you vary one variable, the one you think is responsible for changing the outcome. If your hypothesis is correct, your outcome will change in the way you predict. If not, you have to change your hypothesis and look for other causes.

In drug testing, the "control" group isn't fed any drug at all, they just think they were fed a drug. Experiments with living things are notoriously hard to control, especially with intelligent subjects who will "feel better" based on what they think the drugs "should" do. Sometimes they'll even heal better, since (in a poorly understood way) there is evidence that your state of mind and expectations of getting better will influence your chances of getting better. Control groups can be used to separate that effect (the "placebo" effect) from any actual effects of the drug.

The power of the scientific method isn't so much in that scientists are usually right in their hypotheses, and experiments prove it. It is rather than no matter what the outcome, you learn something more about how things REALLY behave, instead of how you THINK they behave. Eventually you approach the truth, or at least a good enough theory that most of the time the things the theory predicts will happen when you change variables, actually happen.

So whether or not your hypothesis is proved correct by your experiment, if you set it up properly, and carried it out carefully, in a way somebody else could repeat if they followed your technique...then you learn about the true behavior. If nothing else, they learned that the "obvious" variable didn't in fact control the outcome, and need to focus on something else.