Working with Misspecified Regression Models

Conventional statistical modeling in criminology assumes proper model specification. Very strong and unrebutted criticisms have existed for decades. Some respond that although the criticisms are correct, there is for observational data no alternative. In this paper, we provide an alternative.

Saved in:  
Bibliographic Details
Authors: Berk, Richard (Author) ; Brown, Lawrence 1940- (Author) ; Buja, Andreas (Author) ; George, Edward (Author) ; Zhao, Linda (Author)
Format: Electronic Article
Language:English
Published: 2018
In: Journal of quantitative criminology
Year: 2018, Volume: 34, Issue: 3, Pages: 633-655
Online Access: Volltext (Resolving-System)
Journals Online & Print:
Drawer...
Check availability: HBZ Gateway
Keywords:
Reproducibility
Misspecification
Regression
Description
Summary:Conventional statistical modeling in criminology assumes proper model specification. Very strong and unrebutted criticisms have existed for decades. Some respond that although the criticisms are correct, there is for observational data no alternative. In this paper, we provide an alternative.
ISSN:1573-7799
DOI:10.1007/s10940-017-9348-7