Conceptual Model: Causal Model

Conceptual Overview and Discussion

A conceptual model consists of relationships between core concepts. Although a core concept sometimes corresponds to an observable constant, such as “nature” or “organization,” in a conceptual model core concepts refer only to phenomena that can occur in different variations. For example, leadership style is a phenomenon that has different variations, such as “a participatory leadership style” or “an authoritarian leadership style.” Because of this variation, researchers often use the term variables to indicate the core concepts of their research. Some variables, such as “sex,” are nominal variables, and can be presented only in terms of distinct categories, such as “man,” “woman.” Ordinal variables, such as “level of commitment,” can be presented in terms of “more” or “less,” whereas interval variables, for example, “age,” have a variation in which the numeric distance between two ranking points can be fixed.

When we consider the relationship between core concepts, we refer to a causal relationship, in which we differentiate between cause and effect. We recognize a causal relationship in phrases such as “X causes Y,” “X leads to Y,” “the consequence of X is …,” “X influences …,” and so on. Generally speaking, we define a relationship between two variables X and Y as a causal one if we assume that, as a result of a change in X, a change will occur in Y. For example, we expect a causal relationship between a leadership style and the business's profits when we assume that a change from an authoritarian style to a participatory style will cause a change in the volume of the profit.

Labeled according to the nature of the assumed effects, three different types of a conceptual model can be distinguished (Figure 1).

A direct effect conceptual model suggests that we study the immediate influence of the independent[Page 203]variable X (cause) on the dependent variable Y (effect). For example, we expect that intensive smoking (X) causes an increase of lung cancer mortality (Y).

It is possible that the independent variable (X) affects the dependent variable (Y) indirectly, via a mediating variable Z. For example, an increase in the level of economic activity (X) will lead to a higher quality in the healthcare system (Z), and that the latter will increase the population's general state of well-being (Y).

It can also occur that there is a direct effect of variable X on variable Y, but only under particular circumstances. In this case we speak of a moderating effect, because we expect that a third variable Z modifies the direction or the strength of the direct relationship between variable X and variable Y. For example, the level of financial reward (X) will have a direct effect on the level of performance (Y). However, we expect that this relationship will be stronger for young people than for those who are older. Hence, we expect a moderating influence of age (Z). Please note that the moderating variable Z does not have an effect on either X or Y, but only on the relationship between the two variables.

...