Question

Lets say you wanted to predict success in a college course - that is, you wanted to isolate some variables that have a causal influence on good performance in class.
What are some potential INDEPENDENT variables that you might manipulate in an experimental design to test their effect on the DEPENDENT variable of course performance?

Answer 1

In an experiment, the independent variable is manipulated and the effects observed. These observed effects are called dependent variables. They are often the hypothesized outcome of manipulating the independent variable.

A change in the dependent variable depends on the independent variable, hence the name. The dependent variable responds to the independent variable, and it’s this relationship that researchers attempt to measure when conducting experiments.

A well-designed experiment normally incorporate one or two independent variables, with every other possible factor eliminated, or controlled. There may be more than two dependent variables in any experiment.

DEPENDENT VARIABLE EXAMPLES

A researcher might wish to establish the effect of fertilizer on the rate of plant growth; amount of fertilizer is the independent variable. They could regard growth as height, weight, number of fruits produced, or all of these. A whole range of dependent variables arises from one independent variable.

Here, the researchers might also measure other relevant dependent variables which may turn out to be unwanted side effects of the medicine, such as drowsiness.

In any experimental design, the researcher must determine that there is a definite causal link between the independent and dependent variable.This reduces the risk of 'correlation and causation' errors. Controlled variables are used to reduce the possibility of any other factor influencing changes in the dependent variable, known as confounding variables.

In the above plant growth example, the plants must all be given the same amount of water, or this factor could obscure any link between fertilizer and growth.

For the antihistamine trial, a confounding variable may be that a participant’s symptoms could improve simply with the passage of time. This is addressed with a control group that receives no medicine at all, allowing researchers to compare all groups and isolate only the true effects of the medicine. The participants might also be asked to stop all other medication during the experiment – another possible confounding variable.

The relationship between the independent variable and dependent variable is the basis of most statistical tests, which establish whether there is a significant correlation between the two. The results of these tests allow the researcher to accept or reject the null hypothesis, and draw conclusions.