Study Variables

• Study variables are categorized as either dependent or independent. Scientists measure dependent variables to find out how the independent variables have an effect. For example, scientists might measure plant height, number of leaves and their color to find out how different fertilizers and amounts of fertilizers influence plant growth -- in this example, fertilizer type and amount are the independent variables. Scientists might study differences between fertilizers by holding the amount of fertilizer constant, or they might study the effects of fertilizer amount by giving each treatment group a different but constant amount of fertilizer.
  
  

Extraneous Variables

• Extraneous variables are those that might confuse the results of an experiment but are not being tested. Scientists must hold these variables constant so there are no unnecessary sources of variation in the experiment. For example, scientists testing the effect of different fertilizers on plant growth would hold as many other conditions constant as possible by using the same size pots and potting soil, making sure plants receive a constant amount of light and water and starting with healthy plants of approximately the same size and age.
  
  

Control Variables

• Control groups provide an important base line for comparing results, and independent variables must always be held at a constant level for the control group. For example, a scientist studying the effects of fertilizer type on plant growth might give a control group of plants no fertilizer to allow comparison with growth of plants given different types of fertilizer. If the scientist were studying the effect of fertilizer amount, they might decide to use the manufacturer's recommended amount as a control and hold this fertilizer amount constant for the control group of plants while giving other treatment groups other constant levels of fertilizer.
  
  

Random Variation

• Allowing random variation in variables that could confuse the results of an experiment and are not tested can enhance the ability of a study to make predictions. For example, a scientist studying the effects of fertilizer type on growth of indoor plants might want to place groups of plants in different locations with different light and temperature instead of holding these conditions constant, but would hold other conditions -- such as water and potting soil -- constant. In this case, the scientist would need to measure the conditions that weren't held constant, and include them as independent variables in the study that might influence the results.