Using Beta Coefficients to Impute Missing Correlation Coefficients in Meta-Analytic Research: Reasons for Caution

Journal of Applied Psychology, Forthcoming

Fox School of Business Research Paper No. 18-003

50 Pages Posted: 26 Nov 2017 Last revised: 9 Jan 2018

See all articles by Philip L. Roth

Philip L. Roth

Clemson University - Department of Management

Huy Le

University of Texas at San Antonio

In‐Sue Oh

Temple University - Department of Human Resource Management

Chad H. Van Iddekinge

Florida State University - College of Business

Philip Bobko

Gettysburg College - Department of Management

Date Written: November 21, 2017

Abstract

Meta-analysis has become a well-accepted method for synthesizing empirical research about a given phenomenon. Many meta-analyses focus on synthesizing correlations across primary studies, but some primary studies do not report correlations. Peterson and Brown (2005) suggested that researchers could use standardized regression weights (i.e., beta coefficients) to impute missing correlations. Indeed, their beta estimation procedures (BEPs) have been used in meta-analyses in a wide variety of fields. In this study, we evaluated the accuracy of BEPs in meta-analysis. We first examined how use of BEPs might affect results from a published meta-analysis. We then developed a series of Monte Carlo simulations that systematically compared the use of existing correlations (that were not missing) to data sets that incorporated BEPs (that impute missing correlations from corresponding beta coefficients). These simulations estimated (mean population correlation) and SDρ (true standard deviation) across a variety of meta-analytic conditions. Results from both the existing meta-analysis and the Monte Carlo simulations revealed that BEPs were associated with potentially large biases when estimating and even larger biases when estimating SDρ. Using only existing correlations often substantially outperformed use of BEPs and virtually never performed worse than BEPs. Overall, we urge a return to the standard practice of using only existing correlations in meta-analysis.

Keywords: beta estimation procedures, meta-analysis, missing data

Suggested Citation

Roth, Philip L. and Le, Huy and Oh, In-Sue and Van Iddekinge, Chad H. and Bobko, Philip, Using Beta Coefficients to Impute Missing Correlation Coefficients in Meta-Analytic Research: Reasons for Caution (November 21, 2017). Journal of Applied Psychology, Forthcoming ; Fox School of Business Research Paper No. 18-003. Available at SSRN: https://ssrn.com/abstract=3075432

Philip L. Roth

Clemson University - Department of Management ( email )

101 Sirrine Hall
Clemson, SC 29634
United States
864-656-2015 (Phone)

HOME PAGE: http://business.clemson.edu/Managemt/faculty/l3_fac_Roth.html

Huy Le

University of Texas at San Antonio

One UTSA Circle
San Antonio, TX 78249
United States

In-Sue Oh (Contact Author)

Temple University - Department of Human Resource Management ( email )

1801 Liacouras Walk
Alter Hall 343
Philadelphia, PA 19122
United States

HOME PAGE: http://www.fox.temple.edu/mcm_people/in-sue-oh/

Chad H. Van Iddekinge

Florida State University - College of Business ( email )

423 Rovetta Business Building
Tallahassee, FL 32306-1110
United States

Philip Bobko

Gettysburg College - Department of Management ( email )

Campus Box 0395
Glatfelter Hall, Room 413
Gettysburg, PA 17325
United States
717.337.6983 (Phone)

Register to save articles to
your library

Register

Paper statistics

Downloads
70
rank
323,333
Abstract Views
302
PlumX Metrics
!

Under construction: SSRN citations while be offline until July when we will launch a brand new and improved citations service, check here for more details.

For more information