Types of Variables in relation to research design

Independent and Dependent Variables

In general, experiments purposefully change one variable, which is the independent variable. But a variable that changes in direct response to the independent variable is the dependent variable. Say there’s an experiment to test whether changing the position of an ice cube affects its ability to melt. The change in an ice cube’s position represents the independent variable. The result of whether the ice cube melts or not is the dependent variable.

Intervening variables

An intervening variable, sometimes called a mediator variable, is a theoretical variable the researcher uses to explain a cause or connection between other study variables usually dependent and independent ones. They are associations instead of observations. For example, if wealth is the independent variable, and a long life span is a dependent variable, the researcher might hypothesize that access to quality healthcare is the intervening variable that links wealth and life span.

Moderating variables

A moderating or moderator variable changes the relationship between dependent and independent variables by strengthening or weakening the intervening variable’s effect. For example, in a study looking at the relationship between economic status (independent variable) and how frequently people get physical exams from a doctor (dependent variable), age is a moderating variable. That relationship might be weaker in younger individuals and stronger in older individuals.

Constant or Controllable Variable

Sometimes certain characteristics of the objects under scrutiny are deliberately left unchanged. These are known as constant or controlled variables. In the ice cube experiment, one constant or controllable variable could be the size and shape of the cube. By keeping the ice cubes’ sizes and shapes the same, it’s easier to measure the differences between the cubes as they melt after shifting their positions, as they all started out as the same size.

Extraneous variables

Extraneous variables are factors that affect the dependent variable but that the researcher did not originally consider when designing the experiment. These unwanted variables can unintentionally change a study’s results or how a researcher interprets those results. Take, for example, a study assessing whether private tutoring or online courses are more effective at improving students’ Spanish test scores. Extraneous variables that might unintentionally influence the outcome include parental support, prior knowledge of a foreign language or socioeconomic status.

Qualitative variables

Qualitative, or categorical, variables are non-numerical values or groupings. Examples might include eye or hair color. Researchers can further categorize qualitative variables into three types:

• Binary: Variables with only two categories, such as male or female, red or blue.
• Nominal: Variables you can organize in more than two categories that do not follow a particular order. Take, for example, housing types: Single-family home, condominium, tiny home.
• Ordinal: Variables you can organize in more than two categories that follow a particular order. Take, for example, level of satisfaction: Unsatisfied, neutral, satisfied.

Quantitative variables

Quantitative variables are any data sets that involve numbers or amounts. Examples might include height, distance or number of items. Researchers can further categorize quantitative variables into two types:

• Discrete: Any numerical variables you can realistically count, such as the coins in your wallet or the money in your savings account.
• Continuous: Numerical variables that you could never finish counting, such as time.

Composite variables

A composite variable is two or more variables combined to make a more complex variable. Overall health is an example of a composite variable if you use other variables, such as weight, blood pressure and chronic pain, to determine overall health in your experiment.

Confounding variables

A confounding variable is one you did not account for that can disguise another variable’s effects. Confounding variables can invalidate your experiment results by making them biased or suggesting a relationship between variables exists when it does not. For example, if you are studying the relationship between exercise level (independent variable) and body mass index (dependent variable) but do not consider age’s effect on these factors, it becomes a confounding variable that changes your results.