Reasoning: Errors in Causation
By David Roberts
(printable version here)

Often writers argue that there is a causal relationship between two or more things. For example, some might argue that the high turnout of Evangelical voters in 2004 led to President Bush’s victory. Unfortunately, many writers make unwarranted logical leaps when portraying one event as the cause or effect of another. Described below are several kinds of common errors when writers discuss causality.

Correlation does not imply causation.

Many scholarly studies use the word “correlation” to describe two occurrences which happen together. A correlation, however, does not necessarily mean that the two events are logically connected, much less that one caused another. There are two specific errors one should keep in mind when discussing correlating events in academic papers.

1. Alternate Causes. Consider a typical correlation: X correlates with Y. As X increases, so does Y. It does not necessarily follow that X caused Y because some third variable, Z, could have caused them both. A classic example of this situation is ice cream sales and the number of drownings in the summer. The mere fact that ice cream sales increase as drownings increase does not mean that one caused the other. Rather, a third factor, such as hot weather, could have caused both.

2. Reverse Causation. In some cases, one event takes place and shortly after, another takes place. Many times, however, the two events take place at the same time. In this case, rather than X causing Y, Y could have caused X. Some may argue that poor economic conditions are the result of high crime: if there is high crime, businesses won’t invest in expensive equipment that would likely be stolen. It could, however, also be the case that poor economic conditions cause high crime: if there is little employment, people may turn to crime to acquire goods or to alleviate boredom. Perhaps both causal relationships are true. The point is, one cannot infer causality with these facts alone due to the complexity of the relationship.

Post Hoc, Ergo Propter Hoc.

This Latin phrase translates to “After this, therefore because of this.” Sometimes writers assume that because one event takes place before another, it must be the cause of it. While this situation is necessary for causation, it alone is not sufficient. Political science professor Dr. Erkulwater provides an example of what she sometimes sees:

“Tax cuts are good for the economy. Shortly after President Reagan cut personal income taxes in the 1980s, the U.S. economy boomed.”

This statement may or may not be true; the point is, one must have more information than simply the time the two events took place to assert that the former event caused the latter.

Single Cause.

A related problem with establishing causality is that there could be more than one cause of an event. If one looks at data and finds that:

1) X took place before Y, and
2) X rises as Y rises, one may be tempted to assert a causal relationship between the two.

This observation, however, neglects the possibility that a third variable, Z, is also necessary for Y to happen. X alone does not cause Y; rather, X and Z together cause Y. Alone, either X or Z could be insufficient.

For example, sometimes economists use a person’s “willingness to pay” for a good to determine the demand for the good. Demand, however, is also determined by a person’s ability to pay for the good.

Back to 'Analysis and Argument'
Writer's Web | Writing Center | Make an Appointment | Library
Copyright Info