Criteria for Causality - Essay Example

Criteria for Causality I.D. Number: Term and Year Criteria for Causality Explanation of the Three Specific Criteria of CausalityCorrelation Correlation can explain two or more proximal events or facts but it is different than exact causality. When an action directly leads to certain consequences, it can be held as the cause of those consequences. But correlation is a sort of sequential proximity of an action and an event affected by it. Hence, correlation alone is not causality, but can be considered as a determining factor.

(Verma and Pearl, 1990)According to Statistical Assessment Service (2012) under the sponsorship of George Mason University, “In general, we should all be wary of our own bias; we like explanations. The media often concludes a causal relationship among correlated observances when causality was not even considered by the study itself. Without clear reasons to accept causality, we should only accept correlation.” Nonspuriousness A spurious relationship is a mathematical tool which is used to establish that whether two events in correlation have some sort of causal connection or not.

Spuriousness is a property of highly misleading correlations that may appear to be causation. Hence, when a correlation has no spurious relationship, it is a highly possible causation. Hence, nonspuriousness becomes an essential criterion in examining the correlated events and establishing causality. (Pearl, 2000; Verma and Pearl, 1990)Time Order Timing of different events is important since time sequencing of actions and events can determine causation. By measuring time, we can determine when an action took place and when another action or event occurred in correlation or consequence to it.

Hence, time order is also an important tool to inspect causality. Johnson and Bhattacharya (2009, p. 505) point out that in business and economics, “observations are collected in a time sequence with the intention of using regression techniques to predict future trend. In many other experiments, trials are conducted successively in time.” And in any event, a plot of the residuals versus time order can detect a contravention of the postulation of independence among related actions and/or events.

Hence, time order emerges as a criterion of causality.Example of a Causal Statement Active or passive smoking causes respiratory illness due to nicotine. In the above statement, the event of active or passive smoking precedes the event of respiratory illness due to nicotine. The event of respiratory illness due to nicotine does not take place without active or passive smoking. Respiratory illness due to nicotine (even of negligible magnitude) takes place whenever active or passive smoking takes place.

Difference between a Necessary and Sufficient Cause Causes are time and again differentiated into two major categories: (1) Necessary and (2) Sufficient. If event A is a necessary cause of event B, then if event B takes place the event A must also have occurred. In other words, presence of B will mean presence of A as well. But when event A is a sufficient cause of event B, then if event A takes place then event B will also take place. In other words, presence of A will mean presence of B too.

Hence, detection of a necessary cause will help the analyst to track the cause from its consequence. But detection of a sufficient cause will help the analyst to track the consequence of the cause. (Mackie, 1988)Necessary causes are important in finding out the causative, in a time order it will help to understand the event that has already occurred and can further be used in inferential statistics. But sufficient causes are used to predict the consequences of an event; hence detection of sufficient causes is necessary to develop forecasting models.

Examples of necessary cause: Presence of oxygen is a necessary cause for fire to burn. Rotation of earth is a necessary cause for day and night.Examples of sufficient cause: Heat is a sufficient cause for melting of ice. Nuclear explosion is a sufficient cause for radioactive emissions.ReferencesJohnson, R.A. & Bhattacharya, G.K. (2009). Statistics: Principles and Methods. Hoboken: Wiley.Mackie, J.L. (1988). The Cement of the Universe: A Study in Causation. Oxford: Claredon PressPearl, J. (2000).

Causality: Models, Reasoning and Inference. Cambridge: Cambridge University Press.Statistical Assessment Service (2012). STATS.org - What is the difference between causation and correlation? George Mason University, Washington D.C. Retrieved from http://stats.org/in_depth/faq/causation_correlation.htm on 6th August, 2012.Verma, T. & Pearl, J. (1990). Equivalence and Synthesis of Causal Models. Proceedings of the Sixth Conference on Uncertainty in Artificial Intelligence, (July, Cambridge, MA) 220-227.