Which of the following is NOT a criterion for establishing causal relationships?

The correct answer highlights that "randomized controlled trials only" is not a criterion for establishing causal relationships. While randomized controlled trials (RCTs) are a powerful methodological tool for establishing causality due to their ability to control for confounding variables and determine cause-effect relationships, they are not the only way to determine causality.

Causal inference can also be supported through various observational studies, cohort studies, case-control studies, and even through epidemiological models. In epidemiology, establishing a causal relationship often relies on a combination of different criteria and evidence, such as the strength of association, biological plausibility, and the existence of an incremental increase in risk, rather than exclusively depending on RCTs.

Therefore, while RCTs are highly regarded and often considered the gold standard in clinical research, the causal relationship can also be inferred through other means, making it incorrect to state that they are the sole criterion.