Table 3

Impact of the errors of causal effect assignment

True nature of the main exposure effect
CausalNot causal
Reported nature of the studied exposure effectCausalA true causal effect has been discovered. Recommendation to act on this should be considered. Language in the context of a study intended for causal inference.Type I error: there is no causal effect, but it is claimed.
Causal language used or suggestion to take action made when the purpose/ability was to find associations.
Not causalType II error: hiding the true causal objective/result by avoiding use of causal language.No causal language when the objective is prediction or to explore associations.