Types of Variables in Statistics and Research (With FAQs)

Updated September 24, 2022

In research, math, and statistics, a variable is a factor you can measure and change its value over time. As there are many types, it's important you choose a suitable variable when designing experiments, selecting tests, and interpreting results. Learning about variables and how to select one can lead to more accurate calculations, analyses, and results. In this article, we describe the common types of variables and provide helpful answers to questions about them.

Types of variables

Here are the most common types of variables based on classifications by scientists:

Independent variables

An independent variable is a factor that doesn't change its value due to the influence of other variables in your experiment. For example, your age is an independent variable because your address, academic background, experience, or career progress can't change it. Independent variables, or predictor variables, can impact other experiment variables. For example, your age can influence your decision to live in a community for seniors. Another example of an independent variable is the type of lunch you have at work because it can influence your energy levels or productivity.

Dependent variables

A dependent variable relies on other variables to change its value. For example, your rating in an interview is an example of a dependent variable because whether you practised interview questions and dressed appropriately can influence it. A company's growth rate is also a dependent variable because it depends on the strategic business plans developed, the situational analysis conducted, and the performance of employees over a period. Dependent variables, or outcome variables, are typically the focus of experiments and tests. When analyzing the relationship between subjects, scientists work to determine what makes these variables change and measure the expected impact.

Intervening variables

Intervening, or mediating variables, explain the relationship between other experiment factors, typically independent and dependent variables. Unlike other types, they are associations, not observations. For example, suppose you want to predict the life span of the elderly using their wealth as the independent variable. To perform your analysis, you may discover that access to healthcare is an intervening variable connecting wealth to life span. In many experiments, independent variables cause changes in the intervening variable, which impacts the dependent variable.

Control variables

A control variable is a factor that doesn't change during an experiment. As they remain constant, control variables don't affect other variables. Scientists may intentionally include control variables to ensure accurate results. For example, the amount of water and soil used may be control variables in an experiment on plant growth. The time for an experiment is another common variable that researchers use to prevent bias in experiments. Control variables make it easier to repeat experiments, and they can increase the researchers' confidence in experiment outcomes.

Composite variables

A composite variable is a combination of two or more single variables. Combining variables provides more information for statistical analysis or calculations. For example, suppose you intend to make observations about the health of a community. Individually, the weight and height of community members may not provide enough information to determine the community's health. Scientists can use the body mass index (BMI), which is a combination of an individual's weight and height. Composite variables are typically useful for measuring and simplifying complex concepts, such as clinical trials and equipment issues.

Moderating variables

Similar to an intervening factor, moderating variables change the relationship between independent and dependent variables by influencing the intervening variable's effect. It impacts the level, direction, or presence of relationships between intervening variables. For example, suppose you're studying the economic status of individuals by examining how frequently they visit a doctor. In that case, age is the moderating variable because the relationship between these variables may be weaker in young individuals or stronger in seniors. Depending on the experiment, health status, weight, height, stimulus type, and income may also be moderating variables.

Quantitative variables

A quantitative variable, or numeric variable, involves numbers or amounts. You can measure these factors by counting them. For example, your height, paid vacation days, and salary are all quantitative variables. Here are further classifications of quantitative variables according to Statistics Canada:

Discrete variables are numerical variables you can count realistically. As you can determine the number of coins in your wallet, digits in your telephone number, and keys on your keyboard, these are all discrete quantitative variables.
Continuous variables are numerical variables that you could never finish counting. For example, the hair strands on your skin, your height, and your age are all continuous variables because there can have an infinite number of real values within an interval.

Related: What Is Quantitative Analysis?

Qualitative variables

A quantitative variable, or a categorical variable, has non-numerical values. You can typically represent them in groups. For example, eye colour, provinces, and territories are qualitative variables. Here are further classifications of this type of variable according to Statistics Canada:

Nominal variables: describes a name, category, or label without any order. As you may not order your postal code, genotype, or shirt colour, these factors are nominal variables.
Ordinal variables: describes categories, labels, or names that you can order. For example, you can typically order education levels, employee satisfaction, and economic status.

Extraneous variables

An extraneous variable is a factor you're not examining that could affect experiment outcomes or dependent variables. They can unintentionally change how a researcher interprets results. For example, suppose you're observing whether memory capacity relates to interviewing performance. In that case, the candidate's stress level or the time the interview occurred may be extraneous variables.

Confounding variables

Confounding variables are factors you're not examining that could affect both dependent and independent variables of an experiment. A confounding variable is a type of extraneous variable and accounting for it can lead to more accurate experiment results. For example, suppose you're observing a participant's body mass index by examining their exercise level. If you don't consider the participant's age, it becomes a confounding variable.

FAQs about variables

Review the following questions about variables and how to work with them:

How many variables can an experiment have?

Nearly every experiment requires at least one dependent and independent variable. Complex experiments typically have many independent factors. While the dependent variable is the experiment's result, the independent variable is what you test. Other variables affect or impact the dependent and independent variables. For example, many experiments involve controlled factors.

How do you design an experiment with variables?

An experiment's design is a plan to analyze the relationships between variables. Follow these steps to design an experiment:

Determine the subject and question you want to answer.
List all variables, including confounding and extraneous variables.
Develop a theory of what the experiment proves.
Decide how to control the independent variable.
Select the number of study samples or subjects.
Assigns subjects to respective groups.

In what ways can you control confounding variables?

Here are tips to help you control and account for confounding variables:

Adjust study parameters to reduce the effects of these variables.
Compare study groups that contain the same level of confounding variables.
Spread confounding variables randomly between study groups.
Remove samples, study groups, or subjects that contain confounding variables.
Create subgroups where confounding variables don't vary.

How do you identify control variables?

Here are questions to help you identify an experiment's control variables:

What experiment aspects remain constant?
What experiment aspects would you change that would lead to bias?
What important aspects aren't you focusing on during the experiment?
What aspects can you indirectly control through randomization and statistical control?

What is the relationship between a variable's frequency and distribution?

A variable's frequency refers to the number of times its value occurs. Related to frequency, a variable's distribution is its pattern of frequencies. It consists of all possible values of a variable and the frequencies associated with those values. Visit Statistics Canada to learn more about data exploration and variable types, including measures of dispersion and frequency distributions.

What is the difference between a variable and a parameter?

While variables and parameters are similar and important concepts, there are a few differences between them in research and statistics. Variables change during experiments, tests, and studies. In comparison, a parameter is a factor that defines an experiment's population. For example, you may find the average number of cars in a building, which is a parameter of your experiment. Parameters typically remain constant during a study.

Now that you have discovered the different types of variables, you can choose the appropriate one when designing or interpreting your experiment.

Please note that none of the companies listed in this article are affiliated with Indeed.

Types of Variables in Statistics and Research (With FAQs)