A variable is something that can be changed or varied, such as a characteristic or value. Variables are generally used in psychology experiments to determine if changes to one thing result in changes to another.
Variables play a critical role in the psychological research process. By systematically varying some variables and measuring the effects on other variables, researchers can determine if changes to one thing result in changes in something else.
The Dependent and Independent Variables
In a psychology experiment:
- The independent variable is the variable that is controlled and manipulated by the experimenter. For example, in an experiment on the impact of sleep deprivation on test performance, sleep deprivation would be the independent variable.
- The dependent variable is the variable that is measured by the experimenter. In our previous example, the scores on the test performance measure would be the dependent variable.
Extraneous and Confounding Variables
It is important to note that the independent and dependent variables are not the only variables present in many experiments. In some cases, extraneous variables may also play a role. This type of variable is one that may have an impact on the relationship between the independent and dependent variables.
For example, in our previous description of an experiment on the effects of sleep deprivation on test performance, other factors such as age, gender, and academic background may have an impact on the results.
In such cases, the experimenter will note the values of these extraneous variables so this impact on the results can be controlled for.
There are two basic types of extraneous variables:
- Participant Variables: These extraneous variables are related to individual characteristics of each participant that may impact how he or she responds. These factors can include background differences, mood, anxiety, intelligence, awareness and other characteristics that are unique to each person.
- Situational Variables: These extraneous variables are related to things in the environment that may impact how each participant responds. For example, if a participant is taking a test in a chilly room, the temperature would be considered an extraneous variable. Some participants may not be affected by the cold, but others might be distracted or annoyed by the temperature of the room.
In many cases, extraneous variables are controlled for by the experimenter. In the case of participant variables, the experiment might select participants that are the same in background and temperament to ensure that these factors do not interfere with the results. If, however, a variable cannot be controlled for, it becomes what is known as a confounding variable. This type of variable can have an impact on the dependent variable, which can make it difficult to determine if the results are due to the influence of the independent variable, the confounding variable or an interaction of the two.
Operationally Defining a Variable
Before conducting a psychology experiment, it is essential to create firm operational definitions for both the independent variable and dependent variable. An operational definition describes how the variables are measured and defined within the study.
For example, in our imaginary experiment on the effects of sleep deprivation on test performance, we would need to create very specific operational definitions for our two variables. If our hypothesis is "Students who are sleep deprived will score significantly lower on a test," then we would have a few different concepts to define. First, what do we mean by students? In our example, let’s define students as participants enrolled in an introductory university-level psychology course.
Next, we need to operationally define the sleep deprivation variable. In our example, let’s say that sleep deprivation refers to those participants who have had less than five hours of sleep the night before the test.
Finally, we need to create an operational definition for the test variable. For this example, the test variable will be defined as a student’s score on a chapter exam in the introductory psychology course.
Students often report problems with identifying the independent and dependent variables in an experiment. While the task can become more difficult as the complexity of an experiment increases, there are a few questions you can ask when trying to identify a variable.
What is the experimenter manipulating? The things that change, either naturally or through direct manipulation from the experimenter, are generally the independent variables. What is being measured? The dependent variable is the one that the experimenter is measuring.
Evans, AN & Rooney, BJ. Methods in Psychological Research. Thousand Oaks, CA: SAGE Publications; 2014.
Kantowitz, BH, Roediger, HL, & Elmes, DG. Experimental Psychology. Stamfort, CT: Cengage Learning; 2015.
Independent Variables 1
Independent Variables = Grouping Variables
For any research article, be able to correctly identify the independent and dependent variables, and for each independent variable, correctly determine
- the number of groups for that variable,
- whether each independent variable was an active or attribute variable
(Back to top)
Independent Variables: What Groups are Compared?
The first thing you look for in a study is the groups that are compared, which are the levels of the independent variable. Sometimes the title of the study provides a clue about the groups compared.
Example 1: "A nursing intervention to decrease depression in family caregivers of persons with dementia" is the title of an article.
What groups might they be comparing in this study?
If they want to find out if this nursing intervention does decrease depression in this population, they will have to compare caregivers that have received this intervention with those who have not. That difference between the two groups is the independent variable: intervention. We would have to read the article to find out how many levels the variable had, etc., but we know already that this is the variable that defines the groups which form the core of this study.
Example 2: "Gender differences in risk factors for cardiovascular disease."
What groups are likely to be compared? Consider your answer then click here to see the answer.
Example 3: (This is an excerpt from the abstract of a study) Individuals exposed both to cigarette smoke and respiratory pollutants at work incur a greater risk of development of airway hyperresponsiveness (AHR) and accelerated decline in forced expiratory volume in 1 s (FEV1) than that incurred by subjects undergoing each exposure separately. We examined whether smoking cessation or smoking reduction improves AHR and thereby slows down the decline in FEV1 in occupationally exposed workers. METHODS: We examined 165 workers (137 males and 28 females) participating in a smoking cessation programme. Nicotine tablets were used for smoking cessation or smoking reduction. Respiratory symptoms were assessed by questionnaire, FEV1 by spirometry and AHR by methacholine challenge test.
What are the independent variables and dependent variables?
Consider your answer, then click here to see the answer.
(Back to top)
Levels of the Independent Variable.
The independent variable is sometimes called the "grouping" variable, because each group has a specific level or value of the variable. All members of each group will receive or participate in the same intervention, but it will be different for different groups.
The independent variable must have at least two groups or levels, but may have more levels - but usually not more than 5. For example if I want to find out if there are more cars of one color than another at an intersection, I can count the number of blue, red, green, white, and black cars that pass the intersection over the course of a day. The variable is car color. There are 5 levels (blue, red, green, white & black) of the independent variable "car color".
In experimental research, an investigator compares two or more groups that are different on only one factor (or variable), so that any differences between the groups can be attributed to that one difference. The researcher manipulates one variable and measures the effect of that manipulation on another variable.
The term "grouping variable" is especially helpful in studies in which the investigators do not manipulate a variable, but instead compare pre-existing groups. In these studies the subjects are divided into groups on the basis of a particular variable (e.g., those who have had cancer vs those who have not had cancer). Some studies specifically recruit people who fit into one of the groups they wish to compare. In any of these cases, there must be a minimum of two groups - in order to make a comparison. There may be up to 3-5 groups per variable, but usually not more than that.
Example 4: If a study is comparing two types of treatments: a placebo and an experimental treatment, the independent variable is "treatment" and it has two levels: placebo and experimental.
Example 5: If a study is comparing behavior of men and women, the independent variable is gender, and it has two levels: male and female.
Active vs Attribute Independent Variables
If a study is done to determine whether a new drug produces weight loss, and the subjects are assigned to a group who gets a sugar pill or an experimental group who gets the new drug, the independent variable DRUG is an active variable. If a retrospective study is done comparing weight of people at a clinic, and the subjects are divided into two groups according to whether or not they were taking this new drug, DRUG would still be the independent variable, but this time it is an attribute variable, since the type, dose, and duration of taking or not taking the drug was not imposed by the investigators. This type of retrospective study is frequently done to compare the mortality and morbidity of people with different diets and health habits. In these cases, the independent variable is an attribute variable.
However, gender cannot be imposed or changed by investigators, so it is always an attribute independent variable, when men and women are compared in the study (i.e., when it is an independent variable). . Since the levels of the variable are two distinct categories, it is a nominal variable.
A critical characteristic of the independent variable is whether a specifically designed intervention is actively imposed on volunteer subjects by the investigators, or whether the groups were defined by a characteristic of the subjects or chosen by the subjects.
An active independent variable is one that is designed, imposed, controlled by the investigators. This is the highest level of independent variables, met by true experimental studies. It has the advantage of having a consistent intervention, where all subjects receive the same treatment under the same conditions, dose, setting, equipment, frequency and duration of exposure to the variable.
Example 6: A study is done to determine whether a new drug (MiracleX) reduces hypertension more than Norvasc. The experimental group would receive a specific dose, schedule, and duration of taking the tablets; and a control group which would be given identical tablets and instructions, but containing Norvasc rather than MiracleX. Subjects would not know whether their tablets contained Subjects would also be required to meet other conditions of the study that would minimize confounding variables - such as maintaining their current activity level, diet, and other drugs taken. The independent variable DRUG would be an active independent variable.
An attribute independent variable occurs when groups are compared, but the grouping variable cannot be chosen and manipulated by the investigators because it is a characteristic of the subjects themselves.
Example 7: One example of an attribute variable is gender. If a study compared men and women on a dependent variable (e.g., response to this new MiracleX drug), then gender would be an attribute independent variable in that study. The study compared the groups, but the investigators did not - and could not - choose which subjects were men, and which were women ;?).
Example 8: A study compared people who have colon cancer and people who do not have the disease, to see if the groups differed in their diet or family history of colon cancer. Here, cancer status would be the independent variable, with 2 levels: with or without colon cancer. The independent variable is attribute, since the presence or absence of the disease was an intrinsic characteristic of the subjects. The diet and family history of colon cancer would be dependent variables.
Example 9: This study is a randomized controlled trial that compares the efficacy of MiracleX, in reducing blood pressure between a group of people with a history of a myocardial infarction (MI) in addition to hypertension, and a group with hypertension but without a history of MI. What are the independent variables here? Is each one an attribute or active independent variable? Click here to see the answer.
Passive independent variables are similar to attribute independent variables, in that they are not constructed and imposed on subjects. These passive independent variables are not active in that particular study, even though the variable could be active in another study. The investigators design the study in such a way to collect data from subjects who have the levels of the variables of interest. For example, looking again at the independent variable DRUG in example 6; the study could be done in a different way, as follows
Example 10: A retrospective study is done comparing blood pressure of people diagnosed with hypertension. Subjects are given a questionnaire to determine what medications they are taking. The subjects are divided into two groups according to whether or not they were taking MiracleX or Norvasc. Subjects not taking either drug were excluded from the study.
In this study, DRUG would still be the independent variable, but this time it is a passive variable. In this study, the dose, schedule and duration of taking the drug were not fixed by the investigators. In addition, the other potential confounding variables held constant in the study of Example 6, so all of those variables, as well as other unknown ones still confound this independent variable. This type of retrospective study is frequently done to compare the mortality and morbidity of people with different diets and health habits.
Another usage of passive independent variables is seen in studies in which the subjects themselves choose which group to participate in. This could be students selecting a particular section of a course, or ovarian cancer survivors selecting which support group to attend. The groups can be compared, but it is likely that the two groups differ on more than just the group chosen.
Example 11: Two groups of people are chosen for this study. The subjects in the TC group have practiced Tai Chi for at least 10 years. A second group was selected so that the subjects were of similar age, but that had not practiced Tai Chi at all. They found that the TC group had significantly better balance, and fewer falls than did the control group. Here the subjects themselves had chosen to participate or not, in Tai Chi, so the independent variable is passive. This illustrates the difficulty with passive independent variables - it is likely that people who would choose and maintain the practice of Tai Chi would be those able to do and continue the practice, whereas those not able to or interested in the practice would not. So, do the people with better balance practice Tai Chi, or do practitioners of Tai Chi develop better balance? We can't differentiate between these two possibilities from this study.
(back to top)