Wrong direction

The wrong direction fallacy is a logical fallacy where the cause and effect of a situation are reversed. The cause is said to be the effect and vice versa, taking the converse of a statement as true.

Put another way, the established fact is 'if p, then q.' Those guilty of committing this fallacy believe erroneously that 'if q, then p.'

Examples

 * Cancer causes smoking.
 * The increase in AIDS was caused by increased sex education. (The converse is true.)
 * Wheelchairs are dangerous-most people in wheelchairs have been in an accident.
 * God created man.

Related
A related fallacy is assuming that there is causation when the direction of causation (if any at all) is unclear. See Post hoc ergo propter hoc. This fallacy is often found in discussion relating to the social sciences where it is difficult to isolate experimental subjects.

For example:


 * Children who watch a lot of TV are often more violent than their peers. Therefore, watching a lot of TV makes children more violent.

Possibly-however, this doesn't rule out the possibility that children who are more violent (for other reasons) enjoy watching TV more than their peers or that the two have no causation link whatsoever.


 * Recreational drugs (alcohol, cannabis, etc.) are used as self medication for pre-existing psychological conditions.

Do recreational drugs cause psychological conditions? Or, perhaps, do those with the predisposition for those conditions have a higher likelihood of taking recreational drugs and then make their conditions worse because of it?


 * Scientists are less likely to be religious than the general population. This is because being trained in the scientific method makes people less susceptible to religion.

As much as it would be nice if this were true, it may be that those who are less susceptible to religion because of other factors (disillusionment, never being indoctrinated, etc.) makes people more likely to get training in the sciences, or, of course, that there isn't any causation link, just a basic correlation.