In my work on Internet use and religion, one of the recurring questions is whether the analysis I am doing, primarily regression models using observational data, provides evidence that Internet use causes decreased religiosity, or only shows a statistical association between them.