What are all of the reasons that causation should not be inferred from comparative or correlational designs?

In a comparative study design causation should not be inferred first because a relationship can be proved to exist in variables. There can also be significant differences in the groups, but this is not a causal relationship between the variables. In Legette’s study he compared results between to public school systems. One of which is a city public school, the other is a county public school. In the city school they thought ability and effort were more important, where in the county school class environment was more important in music. There is a significant difference in each of these groups, and the environment would be different from an urban school compared to a rural school.
In a comparative study you can also logically establish a causal connection between the independent and dependent variables because of this causation can not be inferred. Again in Legette’s study, the comparing of the two types of schools (county and city). There could be many other variables that come into play when just looking at the differences between city public schools and county public schools. In the study there was a difference in the results between the city schools and the county schools. This would probably have to be broken down even more to determine the real cause of this difference because of all the other possible factors involved (ethnic background, social economical status, school staff, etc).

In correlation studies causation can not be inferred because the direction of possible causation is not clear. Also causation can not be inferred in correlation studies because other variables associated with a particular variable that is not included in the study may affect the relationships and may be causally related. You also have to be careful when looking at the variables and not to infer causation between them, they may imply causation but in reality they do not.