Categorical

For a categorical variable, observations fall into distinct groups. For instance, if we are interested in comparing growth rate of purple finch chicks on Cape Cod vs. the Pioneer Valley.  The location is a categorical variable, and the possible values (the categories) for location are Cape Cod and Pioneer Valley.  Other examples could include life stage (juvenile vs. adult), or reproductive state. Often, categorical variables may only have two possible categories, which means they are binary variables.

Note that sometimes numbers may be used to name categories.  For instance, we may say den 1, 2, 3 when referring to hibernation location of black bears.  In this instance the numbers refer to the specific den rather than a numeric variable.

Quantitative

These are variables that are measured on a numeric scale. (ex. Temperature, body size, height, age). These variables are often broken into the sub-types of discrete and continuous quantitative variables.

Discrete

These are quantitative variables that can only take on specific values, and are often encountered with data that takes the form of counts. If your variable can take the values 1, 2, 3, 4…,  such as if you are counting how many basil plants germinated or how many salamanders use tunnels for migration, this would be discrete.

Continuous

These are quantitative variables that can theoretically take on an infinite number of values. If we are considering the growth rate of chicks as above, the growth rate would be considered continuous.  Similarly, temperature would likely be considered continuous, despite the fact that we round when reporting it. Example: 73.291786 degrees gets rounded to 73.3 degrees.

Role of variables in an investigation

Predictor (Independent)

A predictor variable is a variable that you expect will have an impact on the response variable.

Response (Dependent)

The response variable is expected to change relative to the predictor variable. It is what you are interested in predicting or understanding the behavior of, relative to the predictor variables.

Confounding Variables

A confounding variable is a variable that also influences the response variable, but is usually not influencing the predictor variable.  Confounding variables may make it difficult to evaluate the influence of the predictor variable on the response. When designing an experiment, it is important to minimize the impact of confounding variables as much as possible.

For example, in the first lab, our experiment looked at whether sex of the crab would influence escape speed of the crab.  Sex is the categorical predictor variable and escape speed is the quantitative continuous response variable.  We noted that size of the crab might be a quantitative confounding variable, so we adjusted how we were considering the data accordingly.