What licenses a causal conclusion from a study?

Answer

Random assignment of the treatment

Options - A. Random assignment of the treatment - B. A large number of units in the study - C. A strong association between the explanatory and response variables - D. Blinding of the units and the measurers Why - A. Correct — chance allocation spreads every confounder, named or not, evenly across the groups. - B. Size sharpens the estimate of an association; it does nothing about a confounder that is present at every n. - C. A strong association can be entirely manufactured by a third variable. - D. Blinding removes expectation effects; it does not break the link between a confounder and the treatment.

S21 W1 handout (triangle, confounder examples — shape only); IBS1 §1.4 (experimental design vocabulary, CC BY 4.0); original scenarios