Dependent Variable

A dependent variable is a variable that is explained by one or more other variables, which are referred to as "independent variables." The decision to treat a variable as a dependent variable may also imply a claim that an independent variable does not merely predict this variable but also shapes (i.e. causes) the dependent variable. For example, in a survey studying news consumption, exposure to television news could serve as a dependent variable. Other variables, such as demographic characteristics and interest in public affairs, would serve as the independent variables. These independent variables can be used to predict television news exposure and also may be investigated as to whether they also cause one's exposure level.

Researchers often face challenges in establishing causality based on survey data. In causal inference, the dependent variable indicates an outcome or effect, whereas the independent variable is the cause of the outcome or effect. In order to conclude that the dependent variable is caused by the independent variable, the relationship between the two must meet three criteria. First, the two variables must be correlated. That is, a change in one variable must be accompanied by a change in the other. In the case of a positive correlation, one variable increases as the other increases. In the case of a negative correlation, one variable increases as the other decreases. For example, higher levels of education may be associated with lower levels of television news viewing, and if so, there would be a negative correlation between the two variables. If the two variables are not correlated, then there is no causal relationship between them.

Second, the dependent variable must follow the independent variable in the timing of its occurrence. For example, a researcher who seeks to show that one's level of education influences one's level of television news viewing would need to show that changes in the latter occurred after changes in the former. In some instances, it is relatively easy to ascertain the temporal order of the variables. For instance, if a researcher investigates the relationship between children's academic performance and their parents' education levels, then he or she may be fairly confident in claiming that the former happened after the latter. In other cases, however, the time order is less clear. For example, it may be difficult to determine the temporal ordering of political knowledge and television news viewing.

Third, the observed correlation between the two variables must be genuine—that is, it cannot be explained by other variables. Even if watching television news is positively associated with political knowledge, the relationship may be spurious, from a causal standpoint, if it can be accounted for by another variable, such as political interest. If the positive correlation between television news viewing and political knowledge is due to the fact that the two variables are both positively related to political interest, then the causal relationship may not be valid, and thus is only one of noncausal covariation.

In establishing a causal relationship between a dependent variable and an independent variable, it [Page 189]is not necessary for the independent variable to be the only cause of the dependent variable. In other words, the independent variable can be one of many factors that influence the dependent variable. For example, education levels may influence the amount of television news one consumes even if many other variables (e.g. interest in politics) also affect news watching.

...