Causal Reasoning

Preamble: For the purposes of predicting and controlling our environments, we are generally interested in discerning which events or states of affairs are responsible for bringing about other particular states of affairs. Typically, we’ll conclude that some such factor is the cause of a certain effect.

However, determining causes can be difficult. Some difficulties:

(1) Spurious Correlation

(2) Common Causes

(3) Direction of Causality

So, how should you go about trying to establish causal connections between certain types of events or conditions? [How do we avoid the post hoc fallacy?]

Examples: (1) Fear of Doctors, UFO Abductions

(2) Michael Jordan, Chicago Bulls Championships

(3) Morning Coffee intake, affability of one’s instructor

 

 

 

Mill’s Methods

(For identifying the factors causally responsible for a certain effect)

 

1. The Method of Agreement:

(1) Examine instances in which a certain effect occurs. The more variable these instances are, the better.

(2) Try to identify a factor (or combination of factors) that is present in all of those instances.

The method of agreement helps show that a certain factor (or factors) is necessary for bringing about a certain effect.

One can use the method of agreement to undermine a causal link between an effect and some purported cause by showing that sometimes the effect occurs without the factor.

 

 

2. The Method of Difference

(1) Compare situations in which a putative causal factor is present to ones in which it isn’t. The more similar these situations are in other respects, the better.

(2) Determine whether there is any difference in the observed effect.

The method of difference helps establish that a certain factor is sufficient for bringing about a certain effect.

One can use the method of difference to undermine a causal connection by showing that the presence or absence of the factor makes no difference to the observed effect.

 

 

3. The Joint Method

A combination of the methods of agreement and difference. Compare a variety of situations in which a certain factor is present to similar situations in which that factor is absent. Then show that a certain effect is observed in all and only those instances in which that factor is present.

 

 

4. The Method of Concomitant Variations

Demonstrate that quantitative variations in an effect are systematically related to quantitative variations in a particular factor.

 

 

5. The Method of Residues

Show that all of the factors known to have some influence upon a certain effect cannot explain the observed level of the effect. Then attribute the residual influence to an unexamined cause.

The method of residues is most effective in cases where one cannot directly measure the level of influence a factor has over some effect.