Table of Contents

A research variable is something that can vary or change (or that may be susceptible to variation or change).

For example, the number of people on a typical bus on a typical day varies from bus stop to bus stop, as people constantly hop on and off the bus. Thus, "number of people on the bus" is (a) variable.
You may also consider the gender of people on the bus as another variable. Although people don't change their gender when hoping on and off the bus, the relative proportion of women and men on the bus does vary. Thus, "gender of people on the bus" is also (a) variable.
Temperature also varies during the day, getting warmer or colder as the day progresses. Thus, "temperature" is also (a) variable.

A more formal definition for a variable is something like the following: "A variable is any entity that can take on different values" (Trochim, 20001). Thus, "male" and "female" are the two possible values that the variable "gender" can take, while age can take any reasonable value as, for example, number of years since born.


  • A variable has two or more values or attributes. Eg, gender is a variable with two values, male and female; time is a variable with many possible values, such as 1 minute, 2 minutes, etc. A variable with only one value does not vary and, thus, is a constant. You may end up with a de facto constant if your variable is restricted somehow during your research (eg, you use the variable "gender" in a survey but you only survey women).
  • According to their type of measurement, variables can be classified as qualitative and quantitative.
  • According to the type of values they can take (level of measurement), a variable is considered nominal (or categorical), ordinal, interval and ratio.
  • According to their role in experimental research, a variable is considered independent or dependent.
  • According to their role in correlational research, a variable is considered predictor or criterium.
  • A variable is the basis of statistics-based research. Anything that is or remains constant cannot be subjected to statistical analysis. In fact, many research-based interventions in social settings aim to control the variability of events found significant. In time, if the intervention is successful, this may actually render the original variable a constant, thus making it useless for further research. For example, if social research finds that children who don't go to school end up as criminals in adult life, while children who go to school do not, then this variable (child scholarization, which varies between going and not going to school), can "predict" future criminality. If so, there will also be good reasons to make all children attend school, as this should reduce criminality in the future. Yet, once all children go to school, the original variable does not vary anymore, thus becoming a constant. A constant cannot be used for statistical analysis, thus it cannot be used to predict future events, and, thus, becomes useless for statistics-based research2.
1. TROCHIM William MK (2000). Variables. In, The research methods knowledge base. Atomic Dog Publishing (Cincinnati, USA), 2006 (2nd edition). Retrieved from Research Methods Knowledge Base on 11 September 2010.
+++ Footnotes +++
2. Notice, however, the inherent fallacy of this. A constant can still predict future events (eg that children not going to school end up as criminals), but statistics-based research is unable to appreciate it. In fact, future researchers may conclude that scholarization does not affect criminality because scholarization is not a variable anymore.

Did you understand?

Contributors to this page

Authors / Editors


Unless otherwise stated, the content of this page is licensed under Creative Commons Attribution-ShareAlike 3.0 License