Partially Identified Treatment Effects Under Imperfect Compliance: The Case of Domestic Violence

Abstract

Standard-Nutzungsbedingungen:Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden.Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen.Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. www.econstor.eu The Institute for the Study of Labor (IZA) in Bonn is a local and virtual international research center and a place of communication between science, politics and business. IZA is an independent nonprofit organization supported by Deutsche Post Foundation. The center is associated with the University of Bonn and offers a stimulating research environment through its international network, workshops and conferences, data service, project support, research visits and doctoral program. IZA engages in (i) original and internationally competitive research in all fields of labor economics, (ii) development of policy concepts, and (iii) dissemination of research results and concepts to the interested public. Terms of use: Documents in D I S C U S S I O N P A P E R S E R I E SIZA Discussion Papers often represent preliminary work and are circulated to encourage discussion. Citation of such a paper should account for its provisional character. A revised version may be available directly from the author. During the 1980s a set of randomized experiments were carried out to determine the usefulness of a mandatory arrest policy for domestic assault offenders. The first of these was the Minneapolis Domestic Violence experiment (MDVE), which was carried out in 1981. This paper re-examines the data from the MDVE and uses the recent literature on partial identification to determine the implications for a mandatory arrest policy for domestic assault offenders today. I find support for a mandatory arrest policy for domestic assault offenders, even under a set of minimal assumptions on offender and police behavior. JEL Classification:C9, C14, K42

How Do Right-to-Carry Laws Affect Crime Rates? Coping with Ambiguity Using Bounded-Variation Assumptions

“…Some applications to evaluation of criminal-justice policy include Manski and Nagin (1998), Manski and Pepper (2013), and Siddique (2013). The closest methodological precedent to the present study is the Manski and Pepper (2000) study of monotone instrumental variable assumptions.…”

Statistical modeling under partial identification: Distinguishing three types of identification regions in regression analysis with interval data

“…Further developments in that area include work on corresponding confidence region procedures (see, e.g., [7,42,8]), improvements with respect to computational feasibility (e.g., [11,2,62,53]), and extensions to generalized correction methods for misclassification [40,32]. Selected applications include [44]'s study of income poverty measures based on coarsened survey data, [33]'s use of register data to evaluate the German unemployment compensation reform and [60,41,58]'s investigation of treatment effects in observational studies. In another direction, approaches aiming at "most committed regions" containing at least a prespecified percentage of the data can be found in e.g., [49] or [54].…”

Randomized control trials in an imperfect world

“…Current practice when using data from an RCT with imperfect compliance involves reporting point-identified causal effects such as the ITT or LATE only. However, a more complete picture emerges if estimation of partially identified causal effects were to be carried out on the same data [8].…”

“…Previous work of the author contains a larger number of background references for the material presented here [8].…”

“…Another approach, known as "partial identification," is the estimation of bounds, or the range of values that the causal effect obtained from a perfect RCT might take [2], [4], [6], [7], [8]. This involves replacing the unknown counterfactuals, or treatment effectiveness among those for whom assigned treatment is different from the treatment actually taken, with a range of values these unknown counterfactuals can take.…”