Citation

Namboodiri NK. Am. J. Sociol. 1970; 76(1): 83-102.

Copyright

(Copyright © 1970, University of Chicago Press)

DOI

10.1086/224908

PMID

unavailable

Abstract

A unified treatment of the "before-after" and "after-only" designs and their combinations is given. These designs are viewed as instances of fractional or full replicates of 2 factorials. If these designs are viewed in this way the experimenter has a greater chance of avoiding it falls to which he might otherwise succumb in the interpretation of the results. The paper tries to bring out clearly the type of information each design is capable of providing and under what conditions (assumptions) the results from each experiment can be given useful interpretations.