Premium
Major Alternatives to the Classic Experimental Design
Author(s) -
Greeno Catherine G.
Publication year - 2002
Publication title -
family process
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.011
H-Index - 74
eISSN - 1545-5300
pISSN - 0014-7370
DOI - 10.1111/j.1545-5300.2002.00733.x
Subject(s) - causation , risk analysis (engineering) , work (physics) , point (geometry) , computer science , management science , research design , psychology , engineering ethics , epistemology , data science , sociology , engineering , business , mathematics , social science , mechanical engineering , philosophy , geometry
In summary, giving up any of the three defining features of the Classic Experimental Design has (at least) two important effects. Each concession creates a significant threat to our confidence that any improvements observed were actually attributable to the treatment we are studying. At the same time, each concession opens a door by enhancing, sometimes greatly, the feasibility of conducting work in settings that are the most true to everyday clinical practice. When the work is thoughtfully and responsibly conducted, the alternative designs can contribute to clinical knowledge in a way that is equally important to the contributions made by work that conforms to the stringent requirements of the "Classic" design. In the next article, we will discuss single-time-point designs, along with the important dictum, memorized by most everyone who has taken a course called "Introduction to Research" or "Introduction to Statistics" that correlation is not causation.