Full Citation
Title: When Political Scientists Should Avoid Ratio Regressions
Citation Type: Miscellaneous
Publication Year: 2018
ISBN:
ISSN:
DOI:
NSFID:
PMCID:
PMID:
Abstract: Ratio transformed variables are common in empirical political science research. In this article, we summarize the inference problems that arise when using ratio-transformations to de-scale variables of interest. We argue that existing solutions are ineffective in guarding against a range of inference problems. We argue that researchers should instead include the denominator as a control. We use simulated data to demonstrate our proposed specification outperforms any type of ratio specification in most data generating contexts. 3,522 words main 5,283 words total
Url: https://s18798.pcdn.co/abrahamaldama/wp-content/uploads/sites/6464/2018/09/ratio_manuscript.pdf
User Submitted?: No
Authors: Aldama, Abraham; Bisbee, James
Publisher: NYU Wilf Family Department of Politics
Data Collections: IPUMS International
Topics: Other
Countries: