|
In social science research there is often no clearcut, simple formula for deciding what constitutes a causal relationship. This decision will probably involve a judgement call among several possible criteria.
Intentionality is often ascribed to events purported to be causes. When it is possible to establish intent, this strengthens the argument that a factor is a cause. But this is not yet sufficient proof. The intent may be there, but the intended events may not have occurred, or may have occurred through actions by others. Intent is usually difficult to prove and is subject to interpretation.
Events speak louder than words, and intent is only important if it is related, in a causal chain, to the purported causal events. Establishing intent does not prove that “sanctions caused the children to die”. Instead, one has to ask, “What did they die of? Did the lack of crucial supplies contribute to a higher death rate? Did sanctions reduce access to those supplies?”. If all of this can be shown, evidence of intent can form the first link in a chain of causation.
Moreover, cause is not the same as responsibility, in a legal or political sense. Identification of proximal and distal causes helps in developing a case for what is responsible or who is accountable for a particular action. Too often, those seeking to establish blame for changes in humanitarian conditions make claims of cause without building such a case or providing substantive evidence.
In addition to considering intentionality, one must also be careful to differentiate between correlation and causation in establishing linkages between different variables. There may indeed be a correlation between two variables, but there may not be a direct (or even an indirect) causal linkage. To say that “household food security declined during the period of sanctions” is to suggest that there was some correlation between household food security and sanctions, for that particular case. This is a much weaker relationship than causation which would require demonstration that sanctions actually caused, or contributed to, reduced household food security. After all, just because there are firefighters at the scene of a fire, does not mean that they caused the blaze to occur!
How does one variable relate to another in a causal chain? How does one make decisions about “Which came first, the chicken or the egg?”. These judgements are at the heart of elaborating realistic and useful models. There are several standard criteria used to assist in making these judgements. While no variables meet all these criteria perfectly, one can identify as causal those variables for which the maximum number of these criteria fit in a given situation. These variables have different names in various fields, but the logic behind them is the same. Table 1 presents some criteria of causation from various fields.
Some criteria of causation from the field of epidemiology are explained in Box 1. For the purposes of establishing the extent of the causal relationship between variables, the criteria presented in Box 1 should be considered a checklist for practitioners.
|
