Variables in education research are anything that can have different values or vary across learners. Dependent variables are the outcome variables that you collect data about in research, like learning outcomes. They apply to all research designs: non-experimental and experimental. All measurements used to evaluate or understand learning or a learning environment, such as test scores or attitudes, are dependent variables. Pre-tests and post-tests are dependent variables.
Independent variables represent differences in groups that you think might impact the dependent variables. Independent variables can be fixed, meaning they are manipulated by the researcher, or random, meaning they are pre-determined. Fixed independent variables (e.g., instructional style) are used in experimental designs, and participants must be able to be assigned to one value of the fixed variable (e.g., class instruction is based on lecture or active learning). The researcher manipulates the fixed variable to explore its effect on the dependent variable(s). Random independent variables (e.g., gender or religion) are used in non-experimental designs. These variables are not manipulated, but they can still represent a difference between groups on dependent variable(s). Random variables also include manipulable variables that are not manipulated, like which section of a course a student is in. If you have both fixed and random variables, then you have a quasi-experimental design.
We describe experimental designs primarily by the independent variables. If an experiment has one independent variable, then it is called a one-way design. If it has two independent variables, then it is called a two-way design, and so on. If you had one independent variable that had four levels (i.e., each student could be assigned to one of four groups), it would be called a one-way design with four levels. For more than one independent variable, we describe them as a ___ x ___ design. The number of blanks is the number of independent variables you have. In each blank is the number of levels for the independent variable. For example, for a study with two independent variables, one with two levels and the other with three levels, it would be described as a 2 x 3 design (read as “two by three”). For a study with three independent variables, each with two levels, it would be a 2 x 2 x 2 design (read as “two by two by two”).
Included in these experimental design descriptions is whether participants are exposed to multiple levels of an independent variable. For example, if your independent variable was a teaching method (lecture vs. active learning) and if you taught one unit of a course with lectures and another unit with active learning, then the participants (students) would be exposed to both levels of the independent variable. If participants are exposed to all levels of the independent variables, then it is a within-subjects design. If you taught one group of students with lecture and another group of students with active learning (e.g., two sections of a course) and each group was only exposed to one level of the independent variable, then it is a between-subjects design. You can also have mixed-subjects designs in which the participants are exposed to multiple levels of some independent variables but not all of them. Mixed-subject designs are common when one independent variable is within-subjects and another is between-subjects. For example, say you wanted to know if gender affected efficacy of teaching method, and you taught one unit of a course with lectures and another unit with active learning, this design would be mixed-subjects because each student has a gender (between-subjects), and each student gets both teaching methods (within-subjects).
|Variables||Definition||Example||Type of Question|
|Dependent Variable (DV)||Variable about which outcome data are collected||Test grades, number of forum posts, opinions about online learning||DVs are required for all types of research questions|
|Random Independent Variable (IV)||Variable of the participant that is not manipulated by a researcher||Gender, age, number of courses taken online previously||Random IVs are commonly used in non-experimental designs to answer descriptive and relational questions|
|Fixed Independent Variable (IV)||Variable of the participant that the researcher manipulates||The type of instruction (lecture vs. active) that students receive||Fixed IVs are necessary for experimental designs to answer causal questions|
To view more posts about research design, see a list of topics on the Research Design: Series Introduction.
8 thoughts on “Research Design: Dependent and Independent Variables”
Pingback: Research Design: Series Introduction | Lauren Margulieux
Pingback: Research Design: Qualitative, Quantitative, and Mixed Data | Lauren Margulieux
Pingback: Research Design: Levels of Measurement | Lauren Margulieux
Pingback: Research Design: Survey Design, Demographics, Validity, and Reliability | Lauren Margulieux
Pingback: Research Design: Inferential Statistics for Causal Questions | Lauren Margulieux
Pingback: Research Design: Preparing Data for Quantitative Analysis | Lauren Margulieux
Pingback: Additional Analyses: Interrater Reliability and Demographics | Lauren Margulieux
Pingback: Research Design: Improve Your Design Before Collecting Data | Lauren Margulieux