RT Article
T1 When random assignment fails: Some lessons from the Minneapolis Spouse Abuse Experiment
JF Journal of quantitative criminology
VO 4
IS 3
SP 209
OP 223
A1 Berk, Richard
A2 Smyth, Gordon K.
A2 Sherman, Lawrence W. 1949-
LA English
YR 1988
UL https://krimdok.uni-tuebingen.de/Record/1767142730
AB In this paper, we consider what may be done when researchers anticipate that in the implementation of field experiments, random assignment to experimental and control groups is likely to be flawed. We then reanalyze data from the Minneapolis Spouse Abuse Experiment in a manner that explicitly models violations of random assignment. As anticipated, we find far larger treatment effects than previously reported. The techniques developed should be useful in a wide variety of settings when random assignment is implemented imperfectly.
K1 Selection Bias
K1 Domestic Violence
K1 random assignment
K1 field experiments
DO 10.1007/BF01072450