Citation

Soumerai SB, Penfold RB, Libby AM, Lu CY. Psychiatr. Res. Clin. Pract. 2021; 3(2): 98-101.

Copyright

(Copyright © 2021, American Psychiatric Association Publishing)

DOI

10.1176/appi.prcp.20200039

PMID

36101667

PMCID

PMC9176100

Abstract

In their letter regarding our recent PRCP study (1), Spielmans et al. demonstrate a lack of familiarity with rigorous quasi-experimental research designs. Such designs, however, are essential in studies of health policies, which can seldom or ever be randomized, for example, one can't issue national drug safety warnings to a random sample of the population. Before responding to their specific conclusions, we would like to refer readers to an informative table of the hierarchy of strong and weak research designs. Table 1 is based on 100s of years of science (2). It shows a hierarchy of strong research designs that often yield valid results, in contrast to weak designs without baselines (cited by Spielmans et al.) that are largely untrustworthy (i.e., post-only designs without baselines cannot control for common biases, such as history bias and secular trends)...

[Open access]

Language: en