Evidence for causal mechanisms in social science: recommendations from Woodward’s

Citation: Runhardt, Rosa W (2014) Evidence for causal mechanisms in social science: recommendations from Woodward’s. Philosophy of Science Assoc. 24th Biennial Mtg (RSS)
Internet Archive Scholar (search for fulltext): Evidence for causal mechanisms in social science: recommendations from Woodward’s
Tagged:

Summary

Runhardt discusses how to study the causal mechanism by using the technique of process-tracing, which involves contrasting the observable implications of several alternative mechanisms. This method argues that finding the relationship between a potential cause and effect is not enough, and we should examine other intervening variables between the potential cause and effect. In addition, process-tracing includes both top-down and bottom-up approaches. In this paper, Runhardat focuses on the top-down process-tracing method. The essence of top-down process-tracing method is to contrast rival hypotheses about the causal connection between an independent variable X and a dependent variable Y.

At first, Runhardt discusses Woodward’s theory about causation, which assumes that a successful description about the relationship between cause and effect should refer to causal factors which can lead to change in phenomenon. For example, if there is a variable I which can intervene the variable X and then in turn change Y, then we can confirm the relationship between X and Y by controlling variable I. This means by controlling the intervention variable, we can examine the relationship between cause and effect. There are some requirements for the intervention variable between X and Y. First, I should causes X. Second, X can be only influenced by intervention variable I. Third, any directed path from I to Y should go through X. Fourth, I is statistically independent of any other variables which is not on the I-X-Y path. In addition, the intervention does not actually happen in real case, so we may formulate a hypothetical experiment. Also, it is not necessary that the intervention should relate to human action, it can be a natural process instead.

However, Runhardt thinks that Woodward’s method cannot prove the relationship between cause and effect, because it only focuses on the intervention variables in singular case. Runhardt proposed that we should find an intervention variable by looking at two distinct case studies. That means we should compare two similar cases, in which one has the intervention variable, and another one does not have this variable. Based on the comparison, we can examine the relationship between cause and effect. In other word, we should study a control case and an experimental case, and justify their similarity. Besides, He suggested that moving from singular case studies to general theories is based on a homogeneity assumption, which assumes tat the cause for the effect is the same and provides a basis for comparison.