Assessing threats to the validity of experimental and observational designs.

Experimental designs, particularly the randomized trial, are viewed as the most scientifically rigorous methods by which knowledge is attained. Most clinical disciplines, however, use observational study designs more frequently than experimental designs. Nevertheless, the scientific rigor of experimental designs can be approximated by observational designs if common internal validity threats are recognized and addressed during the development of the research protocol. This paper presents a model of threats that can be applied to both types of designs to compare their scientific merits.