Mediators & Moderators

When a predictor variable and an outcome variable have a significant relationship, which is, in turn, influenced by a third variable, the relationship is said to be mediated by the third variable. In this relationship the predictor variable influences the mediating variable in a causal manner. This mediating variable then leads to the outcome, creating the relationship between the predictor and outcome. It is only because of this mediating variable that a relationship between the predictor and outcome exists. Also, quasi-causal inferences may be drawn from mediated relationships. Mediation is best illustrated by the following model:

As demonstrated by the chart below, numerous studies have examined factors which mediate substance abuse or dependence. In these examples, the predictor variables lead to the mediator which in turn leads to the outcome, which is always substance abuse or dependence. For example, research has found that being raised in a single-parent home can lead to increased exposure to stress and that increased exposure to stress, not being raised in a single-parent home, leads to substance abuse or dependence. The following are some, but by no means all, of the possible mediators of substance abuse.

When a variable indicates the conditions under which a specific effect occurs as well as displays how the direction or strength varies within a given relationship, the variable is said to moderate the relationship. Another explanation is that a moderator variable indicates that an effect only occurs under specific conditions. Unlike a relationship containing a mediator variable, the impact of the predictor variable on the outcome is dependent on the value of the moderating variable. Also unlike a relationship involving mediation, no causal inferences can be drawn from a moderated relationship; relationships can only be described as correlated. However, moderated relationships do identify interaction effects between predictor and moderator variables. Moderation is best illustrated by the following model:

As demonstrated by the chart below, numerous studies have examined factors which moderate substance abuse or dependence. In these examples, the moderator variable impacts the level to which the strength of the relationship varies between a given predictor variable and the outcome of substance abuse or dependence. For example, there is a significant relationship between psychobehavioral risk factors, such as tolerance of deviance, rebelliousness, achievement, perceived drug risk, familism, family church attendance and other factors, and substance abuse and dependence. That relationship is moderated by familism which means that the strength of the relationship is increased or decreased based on the level of familism present in a given individual.

Mediation and moderation research continues to inform the field's knowledge and understanding of a pervasive and dangerous threat to public health, substance abuse and dependence. As the relationships between various predictor variables and the factors which influence them are more closely scrutinized, clinicians and researchers are provided with the necessary information to create more sophisticated and relevant methods of prevention and intervention. While these factors are important to the development of SUDs, there are plenty of other factors both known and unknown that influence the development of this disorder. As such, continued research is both necessary and invaluable.