Specification Curve: Descriptive and Inferential Statistics on All Reasonable Specifications

34 Pages Posted: 25 Nov 2015 Last revised: 3 Feb 2016

Uri Simonsohn

University of Pennsylvania - The Wharton School

Joseph P. Simmons

University of Pennsylvania - The Wharton School

Leif D. Nelson

University of California, Berkeley - Haas School of Business

Date Written: November 24, 2015

Abstract

Empirical results often hinge on data analytic decisions that are simultaneously defensible, arbitrary, and motivated. To mitigate this problem we introduce Specification-Curve Analysis, which consists of three steps: (i) identifying the set of theoretically justified, statistically valid, and non-redundant analytic specifications, (ii) displaying alternative results graphically, allowing the identification of decisions producing different results, and (iii) conducting statistical tests to determine whether as a whole results are inconsistent with the null hypothesis. We illustrate its use by applying it to three published findings. One proves robust, one weak, one not robust at all.

Keywords: Specification Curve; p-hacking

Suggested Citation

Simonsohn, Uri and Simmons, Joseph P. and Nelson, Leif D., Specification Curve: Descriptive and Inferential Statistics on All Reasonable Specifications (November 24, 2015). Available at SSRN: https://ssrn.com/abstract=2694998 or http://dx.doi.org/10.2139/ssrn.2694998

Uri Simonsohn (Contact Author)

University of Pennsylvania - The Wharton School ( email )

3730 Walnut Street
JMHH 500
Philadelphia, PA 19104-6365
United States

Joseph P. Simmons

University of Pennsylvania - The Wharton School ( email )

3733 Spruce Street
Philadelphia, PA 19104-6374
United States

Leif D. Nelson

University of California, Berkeley - Haas School of Business ( email )

545 Student Services Building, #1900
2220 Piedmont Avenue
Berkeley, CA 94720
United States

Paper statistics

Downloads
724
Rank
27,201
Abstract Views
2,553