Branch of statistics concerned with inferring treatment effects
Differential effects play a special role in certain
observational studies in which treatments are not assigned to subjects at random, where differing outcomes may reflect biased assignments rather than effects caused by the treatments.
Definition
For two treatments, differential effects is the effect of applying one treatment in lieu of the other.[1] Differential effects are not immune to differential biases, whose possible consequences are examined by
sensitivity analysis.
Methods
In
statistics and
data science, causality is often tested via
regression analysis. Several methods can be used to distinguish actual differential effects from
spurious correlations. First, the balancing score (namely propensity score) matching method can be implemented for controlling the covariate balance.[2] Second, the
difference-in-differences (DID) method with a parallel trend assumption (2 groups would show a parallel trend if neither of them experienced the treatment effect) is a useful method to reduce the impact of extraneous factors and selection bias.[3] The differential effect of treatments (DET) was explored using several examples and models.
^Vargha-Khadem F, Gadian DG, Watkins KE, Connelly A, Van Paesschen W, Mishkin M (July 1997). "Differential effects of early hippocampal pathology on episodic and semantic memory". Science. 277 (5324): 376–380.
doi:
10.1126/science.277.5324.376.
PMID9219696.
^Nathan BP, Bellosta S, Sanan DA, Weisgraber KH, Mahley RW, Pitas RE (May 1994). "Differential effects of apolipoproteins E3 and E4 on neuronal growth in vitro". Science. 264 (5160): 850–852.
Bibcode:
1994Sci...264..850N.
doi:
10.1126/science.8171342.
PMID8171342.