The example above is typical for a study in which hypotheses are tested. If your research is of a more exploratory nature it may be difficult to draw a causal model. This will be the case when you do not have clear hypotheses to test. It is also difficult to draw a causal model when you are not yet sure how to operationalize the concepts. In this case, you can start with a more generic model, identifying clusters of variables, such as ‘personality’ (rather than ‘extraversion’ or ‘neuroticism’). Figure 9 shows such a model, similar to Figure 8. In Figure 9 the plus and minus signs are omitted because some variables within the same cluster have positive relationships and others have negative relationships with variables in other clusters.
The cluster model is a typical starting point for research seeking to explain a certain dependent variable as good as possible. In this case you try to find the most relevant groups of factors that explain variance in the dependent variable. A typical research question for studies of this kind is “How can Y be explained?” or “Why do people Y”?
If you have trouble thinking about the clusters of variables in your model, the following creative thinking exercise may help. Replace your dependent variable by a different one, and go over your hypotheses again. Would they hold? In the example above, suppose we would replace ‘prosocial behavior’ by ‘aggression’. Which mechanisms would link religiosity with aggression, if there is any relationship to begin with?