Question

The same researcher from the previous line of questions conducts a multivariate regression analysis using the same set of independent variables: parental monitoring, number of delinquent friends, and involvement in school activities. The dependant variable again is the number of times using marijuana.

In the analyses presented below, there is a strong negative association between parental monitoring and the number of times using marijuana. Even with the use of multiple regression analysis, why is not appropriate to conclude that parental monitoring causes marijuana consumption? What two specific reasons or issues would we have to overcome to establish a causal relationship here?

Answer 1

Even though strong negative association between parental control and the number of times using marijuana, issues we have to overcome to establish a causal relationship:

(1) Confounding factor such as high economic condition of parent can be the confounding factor which affects both parental control and the number of times using marijuana. High economic condition makes the parent to concentrate more on their economic problems like business, income tax etc.rather than concentrating on their children. The children also get enough money to spend on marijuana.

(2) In this particular instance, we are analyzing poor parental monitoring causes marijuana consumption. The fact may be the other way. That is: the children might have got involved in marijuana due to bad company. Because of that, they may not be respecting thr parents. Thus parental control may be the effect rather than the cause.